• Uncategorized

Visualization in the Medical Field Using Volume Data and 3D

Visualization in the Medical Field 17

Visualizationin the Medical Field Using Volume Data and 3D

Visualizationin the Medical Field Using Volume Data and 3D

Inthe medical field, visualization of various body parts is veryimportant in understanding the internal functions of variousanatomical structure of the body. The current technologicaladvancements have made possible for the medical data to be put intovisualizations that would help in proper analysis of the interiorbody organs of various organisms across the world (Abrahamsson,Chen, Hajj, Stallinga, Katsov, Wisniewski, &amp Darzacq, 2013).For an accurate diagnosis to take place, there is need for propervisualization of the whole context of the organ which requires stateof art techniques of imaging. The process takes the form ofinspection of multiple images in various angles of analysis.Obtaining of these images can be done through the use variousscanning techniques (Amunts,Lepage, Borgeat, Mohlberg, Dickscheid, Rousseau, &amp Shah, 2013).

Itis through these varied scanning parameters that we see the emergenceof visualization techniques which helps various medical practitionersto know the underlying problems in the human anatomical structure.there are various innovations and technological advancements thathave been postulated in aiding on the profound visualization of themedical data (Araki,Maki, Seki, Sakamaki, Harata, Sakaino, &amp Seo, 2014).These data combining techniques entails the MRI- FMRI, CT-MRI andCT-SPCT which is also known as the 3D-Ultrasound-CT. For these setsof techniques, each of them is purposefully set to be engaged tosuite a certain type of pathology (Bader,Kolb, Weaver, &amp Oxman, 2016).

Thecurrent medical tomographic visualization methods have provided acomprehensive volumetric introspection in the body of the patients asexamined by various doctors. These techniques are used for differentfunctions in the body of patient. These functions include,intraoperative guidance, treatment planning, diagnosis andpostoperative control (Berg,Ott, Klapp, Schwing, Neiteler, Brussee, &amp Kersten, 2013).In the current medical field, there exists variety of imagingmodalities that are appropriate for various functional and anatomicalaspects. in most cases, a patient is taken through a series ofinspections in terms of imaging for the doctor to understand theintegrated perspective of the disease (Binnewijzend,Kuijer, Benedictus, van der Flier, Wink, Wattjes, &amp Barkhof,2013).This move has aided in correct and accurate diagnosis and the bestmedicine to apply for the patient. Additionally, the integrated viewof the disease has enabled the proper examination of the developmentof disease and how best it could be handled in terms of medication.

Inthe conventional world, the doctors mainly depended on thetomographic scans which mainly consisted of the 2D visualizationdata. They therefore had a difficult time in understanding theanatomical structure of the body since they had to mash up the slicesand formulate a 3D visualization in their minds (Borgo,Kehrer, Chung, Maguire, Laramee, Hauser, &amp Chen, 2013).In order to overcome the problem of slice by slice investigation ofthe body, the rendering technique of 3D volumetric data analysis wasbrought operation in order to help in the analysis of the medicalvolumetric data. Initially, one method that dominated thevisualization in the medical field was the notion of surface basedrendering technique which indeed helped in the extraction of the ofthe surface models of the pathological and the anatomical structures(Brodlie,Carpenter, Earnshaw, Gallop, Hubbold, Mumford, &amp Quarendon,2012).

Dueto the increased emergence and growth of the hardware of the computerespecially the graphic processing units GPUs which are programmableand the techniques of direct volume rendering, the 3D volumetric datavisualization was made easier in this field (Brodlie,Osorio, &amp Lopes, 2012).The direct volume rendering technique was substantial in directlyvisualizing a volume dataset without creating an advent ofintermediate representation of the surface. Therefore, the use ofvolume data and 3D visualization has been widely used in the field ofmedicine with keen attention to creating a clear advent of diseasediagnosis (Carrasco-Zevallos,Keller, Viehland, Shen, Waterman, Todorich, &amp Toth, 2016). This research paper gives a critical analysis into the twowell-known techniques of visualization which are the volume data andthe 3D techniques.

Theresearch takes a more specific analysis by narrowing down on themedical field. It is important to understand the best ways in whichmedical practitioners are able to diagnose their patients in the mostaccurate way and give the best advice on the medication. Themilestone seen in the technological advancements in the graphics and3D visualization is applied in the context of medical sector with themain aim of knowing their immense significance (Cavalcanti,Rocha, &amp Vannier, 2014).The research includes the investigations of various researches acrossthe world with a view to bringing out the differences in thesetechniques and their importance to the medical field. More of thisresearch gives a detailed analysis of the tomographic imaging methodsin the medical field, the pipeline for medical visualization, therendering that is accelerated by the hardware and the multi-volumerendering which may be direct of flexible (Cevidanes,Bailey, Tucker, Styner, Phillips, &amp Turvey, 2014).Since these two techniques of visualization goes hand in hand, theirmedical field use is explored concurrently.

MedicalVolume Visualization

Thereexists a variety of aspects of visualization of medical volume. Theseaspects vary with the type of data and visualization technique to beput in place. The detailed information needed from the volume dataalso forms the fundamental basis of their difference. Ideally, theconstructs of the problem can be categorized into three distincthierarchy of abstraction as shown on the graph below.

Chart1.0 [This is the representation of the levels from the developmentsof the whole context of management of visualization volume datasetsanalysis. The interactive platform created here interconnects fromthe creation of the algorithm in the volume rendering to the makingof a visualization applications. The automation of this process isvery essential for creation of efficiency of the whole volumevisualization process.]

Algorithmfor Volume Rendering

Fromthe figure 1, the first step in the medical volume visualization isthe setting up of the algorithm which is essential for creating aninteractive examination into the dataset. This step is deemed to bethe basic rendering technique for the volume dataset. Here, it shouldbe designed in such a way that it is appropriate for medical imageexamination and one should ensure that the visualization permits aninteractive analysis of the sets of medical data in question (Chan,Conti, Salisbury, &amp Blevins, 2013).There are two aspects of the medical data that these this level ofalgorithm must ascertain. It must determine whether the algorithmshould concentrate in the rendering of a specific dataset inparticular way or institute a functionality which would allow forusage of a variety of medical data. This specification is veryimportant as it sets forth the fundamental functionality of the wholesystem of visualization (Chen,Lyons, White, &amp Patel, 2013).It is also at this level that the algorithm will dictate the whetherit is destined for single volume visualization or the multiple volumevisualization. From this perspective, there exists a differentialtrade-off between the rendering performance of the algorithm and itsgenerality (Chen,Kaufman, &amp Yagel, 2012).These sets should therefore well-constructed to permit the high levelof visualization of the medical dataset.

InteractiveVolume Visualization Application

Thesecond step is the building of an interactive medical layer which isdeemed to be on top of the rendering algorithms. It expected thatthese applications should be stipulated to suit the regulations ofvarious medical practices at the same time help the interactiveinvestigations of the whole context of medical volumes data array inthe most intuitive way possible (Chiw,Kindlmann, Reppy, Samuels, &amp Seltzer, 2012).Hence two stratagems are put in place in order to encompasses variousdiagnostics purposes for most practitioners. The visualizationapplication can be made to encompass a generic functionality whichcan address the varied levels of medical examination. On the otherhand, it can be specific to a particular task where the visualizationapplication can be used to solve a particular medical problem andrestricts the communication with other functionalities (Crum,Hartkens, &amp Hill, 2014).

Thesetwo sets of medical data analysis are enshrined in the whole contextof both the 3D and volume visualization. It should be noted as wellthat most of these analyses are geared towards ensuring theanatomical diagnostic processes are well investigated and asubstantial solution brought forward (Dong,Haupert, Hesse, Langer, Gouttenoire, Bousson, &amp Peyrin, 2014).The second step therefore gives much attention to implementation ofthe commands accentuated by the algorithms stipulated in the firststep of the analysis. The first type of functionality which iscreation of a platform of multiple users allows for high level offlexibility (Donner,Menze, Bischof, &amp Langs, 2013).The second type of functionality embraces the fact that there is needfor concentration to some specified areas of concern in the medicalfield and part of the body. For example, in the diagnosis of cancer,it is important to institute a specific algorithm that would help inthe narrowing down on the root cause of cancer (Draenert,Coppenrath, Herzog, Müller, &amp Mueller-Lisse, 2014).


Thethird step now elevates into automating the whole visualizationprocess. This automation is very critical as it provides a basis forfunctionality of the 3D and volume dataset analysis techniques. Inthe medical and clinical practices, medical practitioners usuallyinvestigate similar medical predicaments more frequently (Earnshaw,&amp Wiseman, 2012).For instance, in the diagnosis of cancer, a specific procedure isfollowed to help in understanding the whole context of disease level.These routinely investigations necessitate the automation of theworkflows postulated in the visualization. The establishment of theautomation of these workflows allows for the improvement and supportfor the visual analysis. At one point of medical investigation, thevarious tasks can be done while the computer automated visualizationanalysis is carried out independently (Ebner,Marone, Stampanoni, &amp Wood, 2013).Additionally, the visualization outcome is deemed to be produced in amore standardized way. The standardization allows for easycomparisons of the whole visualization analysis. It should also benoted that most practitioners need high level of efficiency as givenby the 3D and volume dataset analysis technique (Eklund,Dufort, Forsberg, &amp LaConte, 2013).The easy comparison also allows for easy collaboration with variousmedical experts from various fields of operation. Therefore, thevisualization techniques should strive to achieve two sets of goals.The first one would be to ensure the swift and easy initiation ofvisualization automation and secondly, being able to depict thevisualization results in a more intuitive way (Evans,Romeo, Bahrehmand, Agenjo, &amp Blat, 2014).

Overviewof The Thesis

Inthe wake of medical examination, visualization of the patient’sbodies is at the center of understanding the diseases these patientshave. Additionally, visualization allows for clear investigation ofthe body and therefore the inception of an informed decision over thebest treatment to initiate for the disease (Fang,Pouyanfar, Yang, Chen, &amp Iyengar, 2016).In this context, there exists various techniques which the doctorsuse for visualization of the both the external and the anatomicalstructure of the body. These techniques include the volume data andthe 3D. the combination of these techniques are very essential inhelping understand how best the internal organs could be discerned.The role of technology should not be underestimated in this stancesince it gives much attention to the notion of developing a state ofart techniques for visualization of the body structures for medicalpurposes (Fedorov,Beichel, Kalpathy-Cramer, Finet, Fillion-Robin, Pujol, &amp Buatti,2012).

Themain aim of this research paper is to investigate more about thesevisualization techniques and provide some substantial grounds forwhich these techniques are significant in the medical industry. Theresearch in the first instance, investigates into various imagecapture techniques which are in line with the current visualizationtechniques in order to create a comparison within the field ofmedicine. Therefore, the scanning techniques have been discussed inthis research to give way to a proper visualization stances of thebody by the medics. In the recent past, the medical practitionersportend to utilize the non-invasive image capturing techniques tocreate a profound basis for accurate diagnosis (Ferroli,Tringali, Acerbi, Schiariti, Broggi, Aquino, &amp Broggi, 2013).Notably, they have used the scanning techniques that can explicitlyenvisage even the inner structures of the body to the finer details.The congruence with the 3D visualization technique is drawn on theaccount that these scanning techniques are relatively compatible withthe visualization methods which allows for the investigation of thefiner details of the body parts which is essential for diagnosis bythe doctors (Fiolka,Shao, Rego, Davidson, &amp Gustafsson, 2012).

Afterthe introduction in to the imaging, this research investigates intothe first step of volume visualization where it describes thegraphics processing unit (GPU) rendering methods used formulti-volume scenes. Additionally, it describes how the techniqueallows the volume visualization to be presented on a render graph andits importance in the medical field. From the graph formed, there isthe dynamic generation of the of the shaders that would allow for theoptimized rendering of the visualizations of the whole-body part. Themain merits of this level is generality where it can be used forvariety of volumes datasets (Fong,Lamhamedi-Cherradi, Burdett, Ramamoorthy, Lazar, Kasper, &amp Amin,2013).The fact that it could be used in various volume datasets makes iteven more proper for the volume visualization. It also makes it to beapplied to multiple modalities as well as different time schedules.The concept of modulation can furthermore be used in the making avisualization which is single-volume (Fout,&amp Ma, 2012).

Theresearch also discusses how the visualization application creates aninteraction with the algorithms set forth in the first layer ofvolume visualization. In so doing, it digresses on the main pivotalpoints in ensuring a high level of interaction within thevisualization interface. This proposition is strengthened by the factthat these visuals needs to be presented in a manner that depictseven the finer details for patients’ body for proper diagnosis(Gao,Penney, Ma, Gogin, Cathier, Arujuna, &amp Hancock, 2012).

Inthe first instance, there is the visualization gadget that relays thewhole context of the flexibility of render graph to the userdirectly. The proposed users are the medical experts invisualization. Who would need to the service of creation ofsubstantial visualizations for their medical issues at hand. Thepractitioners are expected to take advantage of the flexibility ofthe render graph to investigate into the finer details of the bodystructure in order to understand various effects of disease withinthe body (Garyfallidis,Brett, Amirbekian, Rokem, Van Der Walt, Descoteaux, &ampNimmo-Smith, 2014).Secondly, the application concentrates on the field of cognitiveneuroscience. Here, the simultaneous visualization of the variousfunctional images is presented by the use specialized visualizationtechniques. These are gathered in cognitive analysis and internalreference volume.

Theresearch goes ahead to unveil a GPU based procedure for deformationof volume datasets in the medical field. The techniques couldpossibly be used for surgery simulation which focuses on the directintegration of the volume deformation which is interactive in natureand construes to the visualization methods (Gee,Prager, Treece, Cash, &amp Berman, 2014).Lastly, the paper discusses on the automation of the whole context ofprocess of medical volume visualization. This research uses theexample of a standardized investigation of intracranial aneurysms.The results of the automated interactive results are presented inthis context to accentuate the volume visualization. In theinvestigation, the importance of automation is revealed on theaccount of creation of high level of efficiency in the inculcation ofthe medical procedures. The first instance is the production of theinteractive results for the medical volume dataset. The interactionresults are then uploaded into the web interface which in turn allowsfor parallel investigations in to the whole context of medicalexamination of the body (Gouws,Woods, Millman, Morland, &amp Green, 2015).The parallel examination is essential as allows for comparison of thebody structure from a healthy perspective and a faulty perspective.The third layer of automation of the volume visualization of medicaldataset is therefore essential and is made possible through thisvisualization technique. Throughout the research paper, thediscussion on the 3D technique is well explained along with thevolume visualization analysis of the dataset to create the comparisonbetween them.

Basicsof Medical Visualization and Imaging

Thevisualization for medical volume produces the 2D pictures of thevolumetric datasets got from a patient using a tomographic imagingtool. There are various tomographic methods of imaging which fitsvarious medical purposes. In trying to investigate the visualizationof the volumetric datasets got from the tomographic techniques,variety of algorithms have been developed to provide the analysisinto specialized tasks to allow for concentration into a particularfield of medicine (Gross,Erkal, Lockwood, Chen, &amp Spence, 2014).In this section, more emphasis is laid on the methods and thealgorithm for the volume visualization. in the recent past, themedical practitioners seek to gain insight into the techniques whichwould reveal the finer details of the body of the patient to allowfor more investigations (Grottel,Krone, Müller, Reina, &amp Ertl, 2015).In so doing they tend to use those imaging techniques that helpsunveiling the finer details of the body structures. The techniquesdeveloped also tries to give images that trues to address a specificdisease or problem within the body. Below are some of the tomographictechniques that have been used widely through the medical field tohelp in imaging for various parts of the body.

TheTomographic Imaging Techniques

MagneticResonance Imaging (MRI)

Thistechnique utilizes magnetic knowledge in providing the images ofvarious tissues with the body. It is important to note that varioustissues within the body reacts invariably to various magneticresonance and fields (Heiland,Schulze, Adam, &amp Schmelzle, 2014).Through the property of nuclear magnetic resonance, this techniquecaptures the images in accordance with the type of resonance depictedby each tissue within the body. The scan done through MRI mainlygives the results of the soft tissues such as the brain and themuscles. The main reason for this imaging is because it relies onfluids which contains chemicals that resonates differently withmagnetic field (Hirsch,Wolf, Heinicke, &amp Silva, 2014).The image capture using MRI is as shown below.


Image1.0 [The perpendicular transverse and the horizontal transversesection of the human head as captured through the Magnetic ResonanceImaging. The transformation is done on the account of the movement ofthe tissues brought out in the context MRI tomographic technique. Themain reason why the soft tissues are clearly depicted is because theycontain the elements that are affected by the magnetic attraction]

Theimages above show the internal sections of the head depicted throughMRI which mainly accentuates the alignment of the inner soft tissuesof the brains and muscles. The fundamental basis of working for MRIis the fact that an atomic nucleus usually produces anelectromagnetic signal when magnetic fields stimulated it (Hirsch,Guo, Reiter, Papazoglou, Kroencke, Braun, &amp Sack, 2014).The signals produced by the nuclei can be reconstructed and measuredto form the tomographic slices that explicates through the observedobjects. In the medical field, the hydrogen nuclei are used in thiscase by examining its magnetic resonance. The patient’s body istherefore subjected to a powerful external electromagnetic field(Hirsch,Guo, Reiter, Papazoglou, Kroencke, Braun, &amp Sack, 2014).The subjection would then lead to the alignment of the whole contextof the hydrogen Nuclei which is deemed to be small dipole magnets ineither non-parallel or parallel to the field of magnetic alignment.The interaction between the molecular atoms of the hydrogen and themagnetic field instituted around the body of the patient allows forthe formation of the waves that are then simulated into tomographicimages (Höhne,Fuchs, &amp Pizer, 2012).The resultant affect is the formation of the images of the fine andsoft tissues within the brain on the recipient sheet of reflection.It is from the image formation the fundamentals of visualizationstart making the whole context of both 3D and volume visualizationquite significant.


Thistechnique is done though capturing a series of X-rays from variousplanes of scanning from an object then projecting them backwards intoa 3D volume. The projection is very essential as it helps inunveiling the internal sects of the image both in 2D and 3D stances.One of the disadvantage of this imaging technique is that it exposesthe patients into high level of radiation which could be detrimentalto the body by causing cancer (Holzinger,Dehmer, &amp Jurisica, 2014).Additionally, these pictorials can only be got along the axial planethus permitting the smooth transformation of the angle of scanning.The CT scanner is usually used in emergency cases as its scanningtime is very low. Below is a computer tomography which shows the scanof abdomen depicting an abscess of liver adjacent (Hopp,Baltzer, Dietzel, Kaiser, &amp Ruiter, 2012).


Image1.1 [ The CT scan was developed from various angles to provide anin-depth information for the next step of analysis. Thiscompatibility with 3D visualization analysis is based on the accuracyof the data capture from the initial point of view. The clearvisualization difference is brought about by the notion of clearX-Rays that are intercepted into the area of visualization (abdomen)to give the real-time data on the same.]

CTwas brought to existence by Godfrey in 1967. The method utilizes theX-Rays to capture tomographic images from the body of the patient.The fundamental principle of this technique is that it gives muchattention to the attenuation of the X-Rays along single projectionsof rays from varied directions around the object of examination. Itis from the measurements that the tomographic slice could be produced(Hornung,Wurm, Bennewitz, Stachniss, &amp Burgard, 2013).

FunctionalMagnetic Resonance Imaging (FMRI)

Thistechnique works on the basis of the principle that the changes in theflow of blood within the body can be detected by the functionalmagnetic resonance of the device. It should be noted that theincrement in the flow of the blood in a specific part of the brain istriggered by the activation of the area in question (Huang,Sirinakis, Allgeyer, Schroeder, Duim, Kromann, &amp Lessard, 2016).The activations are deemed to be responding to specific tasks whichare cognitively accentuated by the patient. at the times of scanning,the patient is subjected to some simple activities which may triggerthe movement of the blood to the subject areas. They may therefore beasked about the subject matter or look at a specific image(Idiyatullin,Corum, Nixdorf, &amp Garwood, 2014).This technology of imaging sets forth the notion visualization bothin 3D and Volume dataset analysis. Below are the results of theimaging of the part of the brain using FMRI.

Source:Golestani, N. (2015, August 12). Brain and Language Lab: Neurosciencecenter.Retrieved November 12, 2016, fromhttp://neurocenter.unige.ch/groups/golestani.php

Image1.2 [ The differential investigation is done on the fundamental basisof instigation of the part of the brain to function. The depiction ofthe two sets of brain shows the movement caused as a result of thehormonal transformation in the brain. The changes are instigated bythe prompting the patient to concentrate in watching a movie orgetting involved into a conversation.]


Thisis the computer tomography through the use of a single photonemission. This technique uses the gamma rays in capturing the imagesof the body structure. In this prospect, a radioisotope whichcontains gamma rays is injected into the patient. The gamma camera isthen used to capture the images as lighted up by the gammaradioisotope in the body (Huang,Sirinakis, Allgeyer, Schroeder, Duim, Kromann, &amp Lessard, 2016).This type of imaging technique is usually utilized in instances where3D visualization is required as it depicts the whole round 3Dstructure of the object in question. The figure below shows the SPECTimage of cardiovascular system of the body of a patient.


Image1.3 [ The cardiovascular is captured after the patient is injectedwith the gamma emitting isotope. The comparison is created on theaccount of changes in the flow of blood. The changes in the colorfrom light green to red shows the changes in the flow of blood. Thevisualization is pegged on the intensity of the reaction of theradioactive material injected into the blood of the patient. It isdue to this effect that there is a difference between the image A andB. The concetration of the Isotope is high on the body part B thanthat of body part A.]

TheThree-Dimensional Ultrasound

Inthis method, sound waves with high frequencies is used to produceimages. This technique has an advantage of the fact that smallhardware is used to obtain image in the quickest and simplest waypossible. the costs incurred for this method is also very low ascompared with other techniques such as CT (Idiyatullin,Corum, Nixdorf, &amp Garwood, 2014).However, the image quality is lower than that of CT and MRI at thepoint of image acquisition, which gives an output spatial location ofthe image outputted. The 3D ultrasound scan of a fetus is as shownbelow.

Source:JUM. (2002). Prenatal Diagnosis of Severe Hypospadias With Two- andThree-dimensional Sonography: Journalof Ultrasound in Medicine.Retrieved November 12, 2016, fromhttp://www.jultrasoundmed.org/content/21/12/1423/F2.expansion.html

PositronEmission Technique (PET)

Itis a medical imaging method that utilizes nuclear to show thefunctional processes within the body. A substance known to beradioactive is injected into the body being investigated (Ieiri,Uemura, Konishi, Souzaki, Nagao, Tsutsumi, &amp Tanoue, 2012).This scanning method is mainly used while carrying out investigationinto the specific brain diseases and also in oncology. One of thedisadvantages of this method is that positrons that are released bythe injected substance in the body are destroyed by the electronswithin the body of the patient that could lead to fussy outcome ofthe whole scan (James,&amp Dasarathy, 2014).

Discussionon Imaging Techniques

Fromthe analysis of the type of the imaging techniques above, it evidentthat the MRI and CT are both tomographic imaging methods that allowsfor production of a 3D information about the body of the patient(Johnson,Fain, Schiebler, &amp Nagle, 2013).The fact that these methods have different effects of the patient’sbody makes them to be used for diverse purposes. In so doing, itcomputer tomography helps in the depiction of clear images of theskeletal system of the body which are deemed to be contrasted withthe blood vessels within the body of the patient as well. MagneticResonance imaging on the other hand does not produce clear images ofthe skeletal system of the body as it lack the liquid content in it(Johnson,Fain, Schiebler, &amp Nagle, 2013).the lack of liquid content makes suitable for the CT to be used forthe investigation. MRI can therefore be utilized for the purposes ofdepiction of clear images of the soft tissues which indeed has liquidcontent in them.

MRIis known to give a high clear quality contrast of images of the softtissues which therefore assists in the analysis of most of the fineranatomical structure and functional aspects of the body (Kehrer,&amp Hauser, 2013).

Thisis done through application of the application of various measurementsequences. It is therefore evident that the two tomographic imagingtechniques discussed on this section are complementary in nature.These techniques are therefore used simultaneously to understand theintegrated perception of the particular disease in a body structure.It should however be noted that the inception of CT method has a sideeffect of ionizing X-Rays which indeed provides a leeway forinception of the whole context of marginalization of the volume data(Kelm,Wels, Zhou, Seifert, Suehling, Zheng, &amp Comaniciu, 2013).The ionizing behavior of the X-Rays has the effect of causing cancerwhich is dangerous to the body parts of the patient. The MRI magneticfiends ofn the other hand does not have the side effects of causingcancer to the body. MRI conversely cannot be used by the patient whoalready metallic implants in their body. The metal implanted in thebody may be altered while using the MRI due to the subjection of thepatient’s body to a stringer magnetic field (Kersten-Oertel,Jannin, &amp Collins, 2013).It is therefore important to give much attention to these cautions asone applies them on body for imaging. The agents for comparison canalso be a source of allergic reactions within the body of thepatient.

ThePipeline for Medical Visualization

Theabove discussed tomographic picturing techniques are able to producea 3D image of the body of the patient which is very detailed. Theanalysis and the interpretation of the data from this tomographicimaging techniques is quite integrated and sophisticated. Ittherefore needs a lot of experience and procedure. One of the mainaims of visualization of the techniques in the medical field is toassist in the medical image data analysis which gives insights on thevery detailed connotations of the image in question (Kessler,2014).Therefore, the medical visualization techniques are deemed to adopt aspecific visualization pipeline which indeed helps in creatingconformity to the analysis. The figure below shows the structure ofthe pipeline in steps and a description follows on the composition ofthe structure brought forward.

Table1.0 Visualization Pipeline


Step One


Data Base SimulationSensors

Raw Data

Step Two


Visualization Data

Step Three


Renderable Representation

Step Four


Displayable Image

Fromthe above table, the first step to be done for visualization is toacquire the input data. the data can be obtained from an arbitrarysource of data for example the real-world measurements, the numericalsimulation or a specific database (Kim,Rushmeier, french, Passeri, &amp Tidmarsh, 2014).The second step is the filtering which helps in the transformation ofthe raw data into a visualization data which is deemed to be quiteabstract in nature. The main reason for this transformation isbecause the raw data is not appropriated for direct visualization.the process entails the removal of the uninteresting data anddenoising the whole sample as well. Step three involves the mappingof the medical visualization (Kim,&amp Chung, 2015).Here the visualization data is changed into a geometricrepresentation which is deemed to be quite renderable. The mappingstage also ensure that certain aspects of the data such as thetexture, color and the size are inculcated. The final step which isrendering helps in the production of a displayable image through theuse of the state of art graphic computer techniques(Kruger, Kuzmiak, Lam, Reinecke, Del Rio, &amp Steed, 2013).

Inthe medical field, the medical visualization pipeline can beconstructed to accentuated the whole context of image analysis to itslogical end. The table 1.1 outlines the contents of the pipeline withkeen attention unveiling certain steps which are usually used in thevisualization analysis. This model incorporates certain aspects ofthe general model of visualization pipeline but adds more of thesteps to give it its authenticity in the medical practice(Kumar, Ludlow, Mol, &amp Cevidanes, 2014).The four main phases of medical visualization pipeline are such asthe reprocessing, data acquisition, visual analysis, and lastly,visualization. These stages are as detailed below.

Data Acquisition

Medical Imaging



Reprocessing and Image Analysis







Visual Analysis

Visual Analysis and Interpretation

TheAcquisition of Data

In3-D medical visualization, the data utilized here are obtained bytomographic imaging techniques in the medical field as discussed inthe previous section. The tomographic technique used is dependentupon the type of main use of the medical data obtained(Laha, Sensharma, Schiffbauer, &amp Bowman, 2012).In most cases, the radiologic procedures explicate how certain imagesare to be obtained in accordance with the particular purpose forexample when analyzing a specific disease. In some cases, numerous 3Dimages which have varied MRI sequences or different modalities aretaken in order to obtain a highly detailed view of the body of thepatient (Limaye,2012, October).

Theimaging techniques discussed earlier helps in the generation of atraverse section of a 3D slices of the body part that is scanned. The2D pictorials can be deemed to be 2D grids which ae uniform where thepixels are considered to represent the values of the data at the gridpoints. The expanse accentuated between the two sets of the gridpoint is dependent upon the modalities of the imaging postulated atevery point of investigation (Lin,Wang, Han, Fu, &amp Li, 2012).The distance is also characteristically equal in both the directions.It is through these varied scanning parameters that we see theemergence of visualization techniques which helps various medicalpractitioners to know the underlying problems in the human anatomicalstructure. there are various innovations and technologicaladvancements that have been postulated in aiding on the profoundvisualization of the medical data (Lüsebrink,Wollrab, &amp Speck, 2013).The projection is very essential as it helps in unveiling theinternal sects of the image both in 2D and 3D stances. One of thedisadvantage of this imaging technique is that it exposes thepatients into high level of radiation which could be detrimental tothe body by causing cancer. Additionally, these pictorials can onlybe got along the axial plane thus permitting the smoothtransformation of the angle of scanning. The CT scanner is usuallyused in emergency cases as its scanning time is very low (Maier-Hein,Mountney, Bartoli, Elhawary, Elson, Groch, &amp Stoyanov, 2013).

Thegeometric positioning of the visualization grid point can bedetermined through multiplication of the points of the grid withinthe expanse of y-andx-directions with 2D index of (j,i)on the grid point. Therefore, this data is not explicitly stored. Toallow for proper 3D visualization, the 2D slices are usually gatheredinto 3D volume (Mailly,Aliane, Groenewegen, Haber, &amp Deniau, 2013).In the context, the data elements here are known as the voxels (thevolume elements) and construct a uniform grid. In order to make a 3Dvolume, every voxel is assigned a 3D index (I,j,k).The location of the grid can be calculated by from its index justlike in the 2D and the distances of the voxels in the threedirections postulated. In the event that the distance of the slice isdeemed to be equal to the distances of the pixels, this phenomenon iscalled the Cartesian (Marcellin,Bilgin, Lalgudi, &amp Nadar, 2012).If the distance of the pixel is different from the slice distance onone direction or more, it would be called Uniform. For every 3D grid,there is a connection of six direct neighbors making it have aregular topology throughout the planes of volume data set.

Thecuboid would be formed when the 8 neighboring points on the grid.Interpolation method is used in the calculation of the data valueswithin the grid (Marcus,Harms, Snyder, Jenkinson, Wilson, Glasser, &amp Hodge, 2013).Most of the medical volume datasets are arranged in the form ofuniform grids. Conversely, there exist other type which permitsgreater levels of flexibility through the varying of the spacing,cell types and the topology. Those networks can frequently be foundin numerical reenactments. Rectilinear matrices and curvilinearnetworks resemble uniform frameworks organized lattices with aconsistent topology yet permit a superior adaption to the hiddeninformation (Markelj,Tomaževič, Likar, &amp Pernuš, 2012).In rectilinear matrices, the dividing in a specific heading cancontrast all through the volume yet is consistent between cells withcomparative files. Rather than that, curvilinear frameworks don`tconfine the places of their lattice focuses. Unstructured latticeshave neither predefined geometry nor predefined topology. Thus,vertex directions and vertex network must be put away unequivocally(McConnell,Trägårdh, Amor, Dempster, Reid, &amp Amos, 2016).This builds the capacity measure and muddles information get to andaddition yet gives, on the other hand, a high level of adaptability.Most famous are tetrahedral matrices, which are exclusively gatheredof tetrahedral cells. The transformation of the two-dimensional datainto three-dimensional data is as shown below.

Figure1.0 [The construction of the volume visualization dataset is done bycorrespondingly joining the 2D plane visualizations to suit in thevoxel. The voxel formed the basis for the 3D visualization that isderived from the many tomographic 2D images.]

Theresults of the automated interactive results are presented in thiscontext to accentuate the volume visualization. In the investigation,the importance of automation is revealed on the account of creationof high level of efficiency in the inculcation of the medicalprocedures (Meijering,2012).The first instance is the production of the interactive results forthe medical volume dataset. The interaction results are then uploadedinto the web interface which in turn allows for parallelinvestigations in to the whole context of medical examination of thebody. Initially, one method that dominated the visualization in themedical field was the notion of surface based rendering techniquewhich indeed helped in the extraction of the of the surface models ofthe pathological and the anatomical structures (Mischkowski,Scherer, Ritter, Neugebauer, Keeve, &amp Zöller, 2014).Due to the increased emergence and growth of the hardware of thecomputer especially the graphic processing units GPUs which areprogrammable and the techniques of direct volume rendering, the 3Dvolumetric data visualization was made easier in this field. Thedirect volume rendering technique was substantial in directlyvisualizing a volume dataset without creating an advent ofintermediate representation of the surface. Therefore, the use ofvolume data and 3D visualization has been widely used in the field ofmedicine with keen attention to creating a clear advent of diseasediagnosis (Morgeneyer,Helfen, Mubarak, &amp Hild, 2013).

ImageAnalysis and Reprocessing

Theconcept of image reprocessing and analysis has the main objective ofensuring that there is the enhancement and the analyzation of the rawimage for better visualization. There are three steps involved inthis investigation for the visualization pipeline. The filteringprocess is the first step which characteristically filters the wholecontext of visualization operations. The second step which issegmentation entails the extraction of the pathological or theanatomical structure of the body parts. The last step which is theregistration step helps in the visualization of the datasets in asimultaneous way (Mori,Endo, Nishizawa, Murase, Fujiwara, &amp Tanada, 2014).In the event that the visualization filtering operations areconducted regularly, the application of the registration and thesegmentation depends on the visualization task in question.Additionally, the order in which the organization registration andthe segmentation is done can be reversed at every point in time.


Segmentationdisintegrates a therapeutic volume dataset into anatomical as well asobsessive structures that are pertinent for a particular perceptionundertaking. In this manner, division contains two perspectives. Fromone viewpoint, significant structures must be dependablydistinguished (Morrison,Hollister, Niedner, Mahani, Park, Mehta, &amp Green, 2015).Then again, the state of a fragmented structure ought to beunequivocally determined. In fact, division doles out to each voxel aremarkable label (name) that shows its enrollment to a particularstructure. These labels are put away in an extra volume, a purportedlabeled volume, which has approach degree as the first dataset. Mostdivision strategies take a shot at 2D pictures. They can be connectedto 3D volumes by preparing them cut by cut. Be that as it may, for afew methods 3D partners have been produced (Nelson,Kurhanewicz, Vigneron, Larson, Harzstark, Ferrone, &amp Reed, 2013).Division can be performed physically, self-loader, or completelyprogrammed. In manual segmentation, a client needs to check thevoxels that have a place with a specific structure physically onevery cut of a volume dataset. This strategy is tedious and generallynot complacent for customary application. Interestingly, completelyprogrammed strategies perform division with no client association(Nkenke,Zachow, Benz, Maier, Veit, Kramer, &amp Lell, 2014).Since this is a complex test, most division methodologies areself-loader. They consider that the identification of significantstructures is an abnormal state errand, which is best performed by ahuman, while the outline of the exact state of a structure can bebetter done by a PC. (Self-loader) division methodologies can begathered into four classes:


Pixel-arrangeddivision strategies also take the power of a voxel into record. Apixel is connected to a specific structure if its power exists in aninterim of a lower and upper force limit. For the choice of an edgemore often than not the histogram of a picture or volume is watched.A nearby least regularly speaks to a limit that ideally isolates twotissue sorts (Oliveira,&amp Tavares, 2014).

Regionbased strategies

Forregion based division not just the magnitude of a pixel butadditionally its neighborhood. E.g., area developing begins with oneon the other hand more client selected seed focal points and includesneighboring voxels until the power of a voxel surpasses a clientcharacterized edge (Peng,Bria, Zhou, Iannello, &amp Long, 2014).The edge can again be picked by means of the histogram. Watersheddivision considers a volume or picture as a topographic scene withedges and valleys. The tallness of a voxel is ordinarilycharacterized by its force esteem or its slope extent. This scene isstepwise &quotoverflowed&quot, which prompts to countless regions,purported catchment bassins. At the point when the water levelsurpasses a watershed between two neighboring catchment bassins, theyare consolidated. The calculation must be halted when a fittingdivision of the dataset is accomplished (Peng,Tang, Xiao, Bria, Zhou, Butler, &amp Mitra, 2014).


Edge-basedsegmentation systems attempt to discover constant edges that encasethe envisaged structure. An imperative illustrative of this class islivewiresegmentation. This technique uses Dijkstra`s diagram seek calculationto find a way with negligible cost between two clients-based controlfocuses in a 2D cut image. The cost work between two neighboringpixels relies on upon their powers and on the angle greatness andslope bearing (Peng,Tang, Xiao, Bria, Zhou, Butler, &amp Mitra, (2014).While one of the control focuses is fixed, the client can supplantthe other until the produced edge best fits the objective structure.By over and again including new control focuses the entire limit ofthe objective structure can be resolved. The model parameters(weights) are iteratively adjusted until an ideal fit is achieved,for case in reliance of the picture`s slope greatness (Post,Nielson, &amp Bonneau, 2012).

Advanceall the more, direct volume rendering gives a sort of verifiabledivision by means of exchange capacities. Here division data is notunequivocally produced, but rather is connected on-the-fly amidrendering. This method is point by point in the past areas(Potter, Rosen, &amp Johnson, 2012).


Enrollmentis the way toward finding a spatial change that adjusts onerestorative picture or volume with another medicinal picture orvolume. After registration, it is less demanding to contrast the twodatasets and each other, and it is conceivable to create consolidatedrepresentations(Preim, &amp Botha, 2013).There are three noteworthy application situations for therapeuticpicture enrollment:

• DifferentPoints in Time For some medicinal purposes a patient istomographically imaged at various focuses in time, e.g. to screen thecourse of an infection or the impact of a treatment. Enrollmentpermits here a superior examination of the gained picture datasets,for instance as for tumor development (Provost,Papadacci, Arango, Imbault, Fink, Gennisson, &amp Pernot, 2014).Besides, enrollment can be connected for the coordinating of pre-,intra-and postoperatively taken pictures. This can ease, forinstance, the assume control of preoperative investigation also,arranging comes about into the operation room. With coaligned pre-andpostoperative tomographic filters the achievement of a surgicalmediation can be confirmed.

• MultimodalImaging To show signs of improvement perspective of a malady, apatient is regularly analyzed with a few distinctive imagingmodalities. Before an immediate examination the checks must beenlisted in light of the fact that ordinarily the patient isdistinctively situated in the diverse imaging gadgets. Besides, thedistinctive modalities may appear contrasting imaging blunders(Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016).

• AtlasMatching Often a patient particular dataset is enrolled with a mapbook dataset that speaks to the anatomical normal of a specific bodypart. This permits, e.g., the examination of the patient with thenormal or the simple recognizable proof of specific structures thatare as of now named in the chart book (Raghupathi,&amp Raghupathi, 2014).

Moreoften than not, enlistment is an iterative enhancement fix. In everydevelopment first the change is marginally adjusted, then one of thetwo datasets are changed as needs arises, lastly the nature of theenrollment as for a specific comparability measure is assessed. Inthe event that the similitude is not adequate, the enhancementprocedure is proceeded. Enrollment procedures can be arranged eitherby the utilized change sort or by the connected similitudeestimation. Concerning the change sort, demonstration basedstrategies Model-based separation techniques utilize an underlyingmodel that makes presumptions about size, shape, dim level conveyanceand so forth of the objective structure. This model is iterativelyfitted to the inspected dataset (Rodt,Bartling, Zajaczek, Vafa, Kapapa, Majdani, &amp Kaminsky, 2014).

Adynamic form model or snake is a two-dimensional parametric bend,which is distorted towards the limit of the objective structure byinside and outer strengths. In this manner, a vitality capacity ischaracterized that is made out of an inward vitality, which speaks tothe smoothness of the bend, and an outer vitality, which is gottenfrom the picture`s power qualities and inclination sizes. Byminimizing this vitality work a smooth limit of the objectivestructure can be found. Swell division extends the snake idea to 3D(Saalfeld,Stojnic, Preim, &amp Oeltze, 2016).

Level-setsegmentation systems likewise utilize inner (smoothness) andoutside(picture) imperatives to decide the limit of an objectivestructure. In any case, as opposed to dynamic shape models, the limitis verifiably characterized by a purported level-set capacity, whichis developed under control of a fractional differential condition.Level-set strategies can be connected for 2D pictures and 3D volumes(Schäfer,Forstmann, Neumann, Wharton, Mietke, Bowtell, &amp Turner, 2012).

Dynamicshape models (ASMs) are parameterized depictions of the state ofanatomical structures. They are produced from various referencedatasets by factual examination and depict the primary methods ofshape variety. For division, the mean state of the sought structureis at first put in the objective picture. The subjection would thenlead to the alignment of the whole context of the hydrogen Nucleiwhich is deemed to be small dipole magnets in either non-parallel orparallel to the field of magnetic alignment (Seeram,2015).The interaction between the molecular atoms of the hydrogen and themagnetic field instituted around the body of the patient allows forthe formation of the waves that are then simulated into tomographicimages. The resultant affect is the formation of the images of thefine and soft tissues within the brain on the recipient sheet ofreflection. It is from the image formation the fundamentals ofvisualization start making the whole context of both 3D and volumevisualization quite significant.


Thevisualization stage entails the conversion of the reprocessed 3Dvolume data to the 2D image (Shin,Orton, Collins, Doran, &amp Leach, 2013).This process is duly named volume visualization and hence consists ofthe mapping of the whole data volume to be presented into arenderable state where it is easy to create a 2D projection owing toits presentation. There are three distinct ways in which the volumedata could be visualized. These ways differ a great deal in terms oftheir mapping and also their rendering steps (Shneiderman,Plaisant, &amp Hesse, 2013).The types of volume visualizations are as shown below.

Plane-BasedVolume Visualization

Thistype of visualization is presented on figure 1.3 (a). In this type ofvisualization, the two-dimensional cross sectional slice of thevolume is deemed to be portrayed in form of a 2D image. The slice isusually perpendicularly oriented to the coordinate axis on themeasurement grid of the voxel. In most cases, the modern displaysallow for arbitrary representation of the slices which is aided byhigh level of computer graphics instituted in the voxels(Spottiswoode,Van den Heever, Chang, Engelhardt, Du Plessis, Nicolls, &ampGretschel, 2013).The next step of mapping entails the notion of computation of thepolygon that represents the two-dimensional cut traverse to thebounding box contained in the 3D volume. In this context, the wholemapping is based on staging the whole notion of extraction of the 2Dimage from the 3D volume data in order to get the real tenets of bodystructure in question. The inception of the high level of computergraphic has made the graphical representation of the data both in 2Dand 3D volume to be quite easy. The rendering step entails thealignment and rendering of the polygon in a parallel way to thescreen (Stankovic,Allen, Garcia, Jarvis, &amp Markl, 2014).The intensity values which are deemed to corresponding to the polygonparallel rendering are therefore mapped into the polygon using the 3Dtexturing techniques. The mapping of the intensity values is usuallylinearly mapped into gray values. In typical terms, any type of imageprocessing like contrast improvement could be used in this case aswell.


Figure1.1 [a depiction of the modern plane based medical volumevisualization. The head has been displayed on the context of theperpendicular transverse plane that cuts symmetrically across head.The internal organs are shown along the traverse plane. the volumedataset is able to allow for the triangulation on to the innerdetails of the body part. For example, it is through this traversesection that the understanding of the cerebrum and cerebellum isenunciated.]

Oneof the merits of plane-phased volume visualization is the fact thatit is similar to the conventional slice by slice investigationtechnique of the tomographic images. Therefore, the doctors and otherhealth practitioners are seemingly familiar with the envisaging ofthe volume visualization at hand (Tan,Zeng, Jiao, Wang, Wei, Thiele, &amp Hung, 2016).Additionally, the fact that the data is presented in a 2D way, itwould be much easier to reduces the advents of occlusion whichusually happens at the times when the 3D volume data is projected toproduce a 2D image. Conversely, the 3D structure of the bidy part insquestion is still not displayed and therefore has to be pictured aconstructed in the mind. This notion could be quite hectic for thepractitiones. It should be noted as well that the state of arttechnolology has been developed to carry out the plane representationof the volume datat in a 3D form, while cutting across a specificplane to display the internal structures (Thorwarth,Müller, Pfannenberg, &amp Beyer, 2013).

Therendering done is based on the fact that internal structures need toexplicitly shown while maintaining the 2D representation. In thiscontext, the mixture of 2D and the 3D volume data is quite essentialas it provides the notion of proper display of the data and thereforethe practitioner would not have the struggle of forming the 3Dpicture in the mind. The roles of the computer graphic cannot beunderestimated here as it helps in ensuring that the internalstructures are well represented along a specific plane that iscreates a slice that cuts across the internal structure of the bodypart(Tian, &amp Waller, 2015).For the figure 1.3 (a) a traverse section of the human head isexamined where the slice is extracted from a particular plane thatmakes the cut symmetrical. This representation is plane based at thesame time it represents a 3d volume data allowing for easyinvestigations into the inner parts of the human head from the brainto the skull.

Surface-BasedVolume Visualization

Thisis representation, the fundamental step here is the making of apolygonal 3D model which is surface based. It is more of tern thannot a triangular mesh of a particular pathological or anatomicalstructure of the body. The pathological or anatomical structure isextracted through the process called mapping. After the extractionthrough the mapping method, the structure is then rendered throughthe use of the 3D rendering techniques. the generation of the surfacemade model is usually done through two ways which are along with thedemarcation of the pre-segmented structure or along the Iso-surface.An iso-surface is that surface in a 3D volume data for which thecorresponding intensity has the same value in every stance(Tobon-Gomez,De Craene, Mcleod, Tautz, Shi, Hennemuth, &amp Lutz, 2013).

Thiscalculation forms the cuboid cells of the volume framework insuccessive request. For each of the eight vertex estimations of amatrix cell it is checked in the event that it lies inside or outsidethe isosurface.There are 256 unique arrangements of vertex states which can becompressed to 15 distinct cases. For each of these cases exists arelating triangle arrangement that approximates the Iso-surfaceinside the cell. These setups are put away and are added to thecreated surface work when a relating vertex arrangement happens. Theposition of the triangle vertices along the cell edges is processedby direct interjection. The walking solid shapes calculation canlikewise be connected to fragmented volume information however thenpurported stair-case curios happen because of the lofty move betweenvoxels inside and outside a divided locale. To smother these antiquesthe portioned volume dataset and additionally the produced work canbe smoothed.

Aproduced triangle work can be directly rendered with traditionalillustrations equipment bolster. To accomplish a three-dimensionalimpression of the structure the surface must be furthermore lit upand shaded. Therefore, the surface-based volume rendering method isfrequently alluded to as surface shaded show (Tomar,&amp Agarwal, 2013).

Thebenefit of surface-based rendering is its immediate support by the 3Drepresentation equipment. Be that as it may, consequently the era ofa surface mesh is a costly pre-handling step. Besides, a surfacemodel displays stand out certain structure of a dataset and not thedataset in general. At the point when diverse structures of a datasetought to be examined all the while, a few surface cross sections mustbe produced and rendered in mix. The diagram below the results of therepresentation of the surface-based visualization Varallyay,Nesbit, Fu, Gahramanov, Moloney, Earl, &amp Neuwelt, 2013).

Source:Juhasz, Z. (2015, March 30). A GPU-based Simultaneous Real-Time EEGProcessing and … Retrieved November 13, 2016,fromhttps://www.researchgate.net/publication/278675948_A_GPU-based_Simultaneous_Real-Time_EEG_Processing_and_Visualization_System_for_ Brain_Imaging_Applications

Figure1.3 [ Only the superficial structures are revealed here to give moredetailed analysis of the same. The real-time processing is done onthe account that the volume visualization is done on a rotationalway.]

Directvolume visualization

Inthe direct volume visualization, there is no use of any intermediatedepiction for rendering. However, there is the direct access of theoriginal volume data set. Essentially, every pixel of a perceptionalray is directed towards the volume data that concentrates the colordistributions at various equidistant points (Velten,Wu, Jarabo, Masia, Barsi, Joshi, &amp Raskar, 2013).The determination of the data color distribution and spatialcontribution is done by a transfer function that is stipulated to mapthe intensity value at the various sample points with the voxel intoan opacity and color value. By the notion of the variation of thewhole context of transfer functions, various structures contained inthe dataset can be easily suppressed or emphasized. In direct volumevisualization, the rendering and mapping helps in the building a unitthat is too invincible to be separated into various independent steps(Volonté,Pugin, Buchs, Spaltenstein, Hagen, Ratib, &amp Morel, 2013).

Oneof the advantages of this volume visualization is the fact that it isable to highlight distinct structures of the volume datasetregardless of the need of a costly reprocessing. Additionally, thereexist numerous shading and rendering methods in which various aspectsof dataset could be emphasized. Clipping methods could be used togive insights to the inner structures. The outer shape on the otherhand is deemed to be provided in simultaneous way on the context.Conversely, this type of visualization is perceived to be quiteexpensive. It is usually deemed to be quite expensive since morecomputer graphics is required which would need more expertise andhence high costs incurred (Wang,Chen-Wiegart, Eng, Shen, &amp Wang, 2016).This makes it more expensive than the surface based volumevisualization. However, it should be noted that there are numerousgraphic processing unit based rendering algorithms which can be usedto achieve interactive rates of frames for the mean-sized volumedatasets in the medical field. Hence this type of volumevisualization is currently used for medical practice. Arepresentation of volume dataset using the direct volumevisualization is as shown in the below diagram.


Figure1.4 [ The direct volume visualization reveals both the superficialand the internal medical volume dataset. The rendering is done onvarious planes in order come up with the whole context of this bodypart analysis. The visualization of the head volume dataset abovereveals that this technique encompasses both the surface based andvolume based visualization in a simultaneous stance.]


Thelast in the medical visualization pipeline is the notion ofanalyzation of the visual brought out and its interpretation. Thedisplayed content of the data is put into a thorough analysis inorder to understand some of the examination about it for example theinvestigation in to the growth of the disease and the extent in whichit has affected the body (Ware,2012). Thedisplayed data is given much thought on the account its role intrying to understand the whole context of the disease in question.The task of visual analysis and investigation is mainly done by theend user of the application in place (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016).Due to the complexity of the volume dataset visual analysis, the enduser interaction has been made in an intuitive way such that itcreates a visual application with a state of art interactionfunctionality.

Thefunctionality is very essential as it helps in ensuring efficiency inthe visual analysis. the visual analysis is based on the fundamentaldifferences found between the visual structure of the patient and thenormally required structure. The application is therefore fitted withthe control visualization which helps in creation of comparison andcontrasting various facets of the body (Würslin,Schmidt, Martirosian, Brendle, Boss, Schwenzer, &amp Stegger, 2013).On the account of the visualization the role of the control visualsshould not be underrated since they give directions on what is reallywrong about the structure at hand. it should be noted as well thatmost of the discrepancies is seen when the real visuals of thepatient are compared with the those of the normal person on the samepart of the body under investigation (Mischkowski,Scherer, Ritter, Neugebauer, Keeve, &amp Zöller, 2014).For example, for the physician to understand the extent of the breastcancer, they would capture the structure of the breast using variousimaging techniques mentioned earlier in this paper. They will thenprocess and reprocess to get to convert it into a 3D volume dataset.The 3D volume data set would then be compared with the inbuiltstructure of the breast in the interactive platform of theapplication in place. This comparison will give the practitioner aneasy time to note some of the effects of cancer on the visualizedstructure of the breast volume dataset (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016).This comparison unveil in the visual analysis is as shown on thediagram below.

Source:Viehland,C., Keller, B., Carrasco-Zevallos, O. M., Nankivil, D., Shen, L.,Mangalesh, S., … &amp Izatt, J. A. (2016). Enhanced volumetricvisualization for real time 4D intraoperative ophthalmic swept-sourceOCT. Biomedicaloptics express,7(5),1815-1829.

GraphicsProcessing Unit (GPU) Deformation of A Medical Volume VisualizationDataset.

Inthe context of volume visualization technique, the technique ofdeformation is widely used in understanding the effects of fractures.The formation of various combinations of the whole context of volumedataset is based on the constructs of the plane, the angles and thematrix of the grid field. The changing of the various points on thegrid helps in the understanding of some vital effects of changes inthe body structure as impacted by the external forces such as theaccidents and various facets of deformation. Even at the points offractures, it is important to understand some of the vital aspects ofthe grid distances that when changed could result into desireddeformation expected when a body of a patient is fractured (Xia,Wang, &amp He, 2013).

Forthe investigation of the GPU-based approach, a volume deformationpipeline was brought into existence. The pipeline consisted of fivedistinct steps which are indeed applied sequentially. The usagedepends on OpenGL and GLSL and can be incorporated effortlessly intothe multi-volume rendering system introduced in Chapter 3. Everypipeline step is acknowledged by a particular GLSL shader. Theseshaders essentially follow up on three information structures (3Dsurfaces): the first volume dataset, the forward disfigurement matrixand the reverse deformation framework (Mischkowski,Scherer, Ritter, Neugebauer, Keeve, &amp Zöller, 2014). Forping-pong rendering the forward disfigurement framework is copied. Inthe event that the disfigurement framework has a lower determinationthan the first volume dataset, an extra low-determination volumedataset is created by mipmapping. This low-determination dataset isutilized for the assurance of the deformation imperatives (Yu,Efstathiou, Isenberg, &amp Isenberg, 2012). The informationstructures are either bound as information surfaces (read) or asrender targets (compose). In pipeline steps where the forwarddisfigurement matrix is utilized for both, perusing and writing, oneof the two twisting network surfaces is bound as information surface,and the other one is bound as render target. The parts of the twosurfaces are changed after every render pass. In the accompanying thefive pipeline steps are clarified in detail.


Inthe control step, the present mouse development is mapped intostandardized 3D demonstrate space, and it is checked if a crash withthe volume happens. For this reason, a solitary section is rendered,and the connected shader is casting a beam from the begin to the endposition of the mouse development. At every examining position, thetwisted surface arrangement is turned upward in the conversedeformation lattice surface (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016). At that point, thecomparing test phase is perused from the first volume informationset, and the related alpha connotation is turned upward in theexchange work table. In the event that the alpha value is moresignificant than zero, an impact is realized. For this situation, thedeformation procedure is introduced by moving the eight frameworkcomponents of the forward distortion matrix that encompass the impactposition equally to the mouse move (Zhao, &amp Xie, 2013).


TheChain-Mail deformation method is acknowledged in numerous passes. Inevery pass the forward distortion lattice surfaces for perusing andcomposing are exchanged. The present position of a network componentis put away in the x-, y-, and z-segment of the related texel. Thew-segment is utilized to store the data if a network component wascurrently moved. This data is utilized to assess if any neighbor of aresearched element was moved in the first pass. If not, the componentis specifically disposed of. Otherwise, it is tried if the presentcomponent must be moved to fulfill the chainmail limitations inrespect to a moved neighbor (Mischkowski,Scherer, Ritter, Neugebauer, Keeve, &amp Zöller, 2014).

Thechainmail deformation is ceased when no more components arepostulated in the present pass. This is verified by equipment upheldimpediment questions which give the number of processed elements(fragments) in the current render pass. When all elements arediscarded, the occlusion query becomes zero and the deformation canbe stopped. The deformation step can be optimized by exploiting thefact that it always starts with eight neighboring grid elements andthat the region of possibly affected elements expands in a singlepass about one along its six border faces. Thus, it is sufficient toinvestigate in each render pass only those elements that can bepotentially moved, and to increase the region of investigatedelements accordingly afterwards (Yu, Efstathiou, Isenberg, &ampIsenberg, 2012).


Therelaxation step is performed in multiple passes as well. In each passfor each element the displacement is computed due to the relaxationrule. If the magnitude of the displacement is too small, the elementis not moved and discarded instead. The relaxation is terminated whena per-pass occlusion query returns that no more element wasprocessed. For performance reasons, in each step only those elementsare investigated that can be possibly moved. This is the case eitherif any of an element’s neighbors or if the element itself was movedin the previous pass. In the first relaxation pass all elements thathave been replaced sometimes during the deformation step are regardedas moved. Like for deformation, the region of possibly affectedelements has to be increased in each pass by one along each border(Velten,Wu, Jarabo, Masia, Barsi, Joshi, &amp Raskar, 2013).


Theinverse displacement position of a forward grid element can becomputed independently of the inverse displacement positions of theother elements. Thus, the grid inversion can be performed in a singlerendering pass, while the iteration loop is placed inside the appliedshader. Thereby, the forward deformation grid texture which wasrecently used for writing is bound as input texture and the inversedeformation grid texture is set as render target.


Fordirect volume visualization of the deformed volume with themulti-volume rendering framework the Scene Node, namely the sub-nodethat is responsible for the volumes, is slightly adapted. Instead ofdirectly looking up the volume sample at the current texturecoordinate, the deformed texture coordinate is first looked up in theinverse deformation grid. Then, the sample value of the deformedvolume is looked up in the original volume texture (Yu, Efstathiou,Isenberg, &amp Isenberg, 2012). Gradients of the deformed volume arecomputed on the fly by finite differences. Since the Scene Node makesthe deformation process transparent for subsequent nodes, volumedeformation can be combined with any existing or newly implementedrender node. Additionally, it is possible to apply either the raycasting or the slice-based rendering technique on it. The figuresbelow show how deformation can be applied when investigating thehuman head.

[Theabove figures are deformed human head scan captured through computertomography technique. The original human head as connoted by the CTresults is (225x256x256). It has then been subjected to variousdeformation grid sizes and algorithms of a) 32 by 32 by 32, throughthe use of Chain Mail only. B) 32 by 32 by 32, through the use ofChain Mail and Relaxation. c)128 by 128 by 128, through the use ofChain Mail, Illumination and Relaxation.]

ModernDevelopment of Solutions Medical Volume Visualization

Theprevious discussions mostly border on the techniques of visualizationof volume dataset and how best we can create a link between theimaging techniques and the visualization techniques. There istherefore the need to understand some of the vital aspects of thesetechniques and how they could be properly inculcated into the medicalfield. There techniques and tools for automated and iterative volumevisualization postulate distinct phases of a frequentativedevelopment process in the field of medical volume visualizationexplanations (Yu, Efstathiou, Isenberg, &amp Isenberg, 2012). Due tothe fact that most of the visualization techniques were put intopractice using the most generic way, they can be adapted for medicalpurposes with a lot of ease. This section proposes a common processin which the iterative development of the volume visualization in themedical field could be brought to existence (Velten,Wu, Jarabo, Masia, Barsi, Joshi, &amp Raskar, 2013). This notion isstrengthened by the fact that the development proposed is based onthe fact that it embraces and allows for the inculcation of the stateof art technology in the whole process of visualization analysis. Themedical volume visualization represents the basis of in depthanalysis of the various facets of the pathological and anatomicalbody structure investigation. By creation of this iterative process,it provides a platform for creation of an application ofvisualization which allow for not only the intuitive interaction withthe end user but also permits intensive and easy investigation in tothese anatomical structures through the use of medical volumevisualization techniques.

FourStages of Medical Volume Visualization

Thereare four stages of Medical volume visualization that are discussed inthis section. The first step is to analyze the 2D data that isacquired through the various tomographic imaging techniques. Thisstage goes along way with the generation of a 3D prototype whichhelps in the investigation of the internal structures as well(Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016). This step isfollowed by the particular interactive application in thevisualization stance then finally ends with a computerizedvisualization service which is expected to be applicable to thelarger amounts of users. The table below shows the visualrepresentation of the steps in which are followed in the iterativevolume visualization development process.

Classical2D analysis

Theprocess of developing the medical volume visualization appropriationusually begins with a particular application module in which the 3Dvisualization methods should be utilized. This concept can beutilized in various aspects of the visualization in the medical field(Laha, Sensharma, Schiffbauer, &amp Bowman, 2012). The analysis canfor instance be used in the examination of a specific disease whichindeed has the effect of redistributing the knowledge of its use andexistence in various sets of investigations. For this applicationcase, ordinarily a work process has been built up that contains thesecuring of the tomographic pictures and the examination of theinformation in a traditional cut by-cut way (Velten,Wu, Jarabo, Masia, Barsi, Joshi, &amp Raskar, 2013). In this way,the included therapeutic specialists or scientists have a reasonablethought regarding which pictures ought to be brought with a specificimaging methodology and which data can be extricated from thisinformation. 3D volume representation procedures can bolster theexamination of this information in a few ways (Laha, Sensharma,Schiffbauer, &amp Bowman, 2012). From one perspective, 3D perceptionfacilitates the examination of three-dimensional information, inlight of the fact that the spatial relationship between variousstructures is straightforwardly shown. In this manner, theexamination procedure can be quickened. Moreover, 3D perception givesnew bits of knowledge into the information and permits theapplication of complex association strategies.

Visualizationof the prototype in 3D

Inthe principal phase of 3D representation, a prototypic volumeperception answer for the obtained picture information is created.For this reason, the nonexclusive volume perception device that wasexhibited in previous sections can be utilized (Bader, Kolb, Weaver,&amp Oxman, 2016). This device permits the intuitive control of therender diagram and, in this manner, its application for variousperception purposes. At initial, a restorative master, who knowsabout the therapeutic application case, also, a representationmaster, who knows the bland volume perception device, investigatewhich render hubs must be consolidated to accomplish the coveted 3Dpresentation of the information. In the event that essential, newrender hubs are executed and incorporated into the framework likeportrayed in Sections above (Laha, Sensharma, Schiffbauer, &ampBowman, 2012). At the point when a representation is found that bestfits to the information, the relating render diagram can beserialized and put away as a XML petition for later reuse.

Contingentupon the application undertaking, a few render diagrams for variousblends of the obtained datasets can be organized. In the applicationand approval stage the predefined render diagrams are connected tovarious chose instances of clinical practice or research.Subsequently, it can be assessed on the off chance that the createdperception designs meet the prerequisites of the medicinalundertaking. For the most part, the render diagrams will beiteratively adjusted and reached out until the restorativespecialists are happy with the outcomes (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016). Advance on, itcould happen that the prototypic 3D representations demonstrate thatthe procured medicinal pictures don`t give sufficient data forperception. For this situation, it ought to be come back to the pastadvancement stage to essentially adjust the picture obtaining workprocess.

Particular3D Visualization Application

Inthe following advancement arrange a tweaked volume perceptionapplication is produced that depends on the representation designsthat were composed in the previous model stage. This application canuse the multi-volume rendering structure what`s more, consolidate itwith an undertaking particular UI. The point is to produce aperception instrument that backings scientists and medicinalspecialists in their day by day work. Therefore, the fundamentalinsights about the utilized render charts ought to be covered up, andcollaboration components ought to be given that fit to theapplication area (Laha, Sensharma, Schiffbauer, &amp Bowman, 2012).Accordingly, the apparatus ought to permit the adaptableinvestigation of standard and non-standard cases. At the point whenfundamental, the hidden rendering structure can be reached out by newrendering procedures that allow a superior improvement of theperception procedure. The assignment specific 3D observationapparatus is connected to all cases that happen in clinical butcorrelates into the medical schedule. Since the apparatus gives anarea specific UI, it can be utilized by analysts and restorativespecialists who have involvement with the therapeutic applicationhowever, they may not be acquainted with the fundamentalrepresentation methods. Accordingly, best-hone representation workprocesses can be expounded that bolster the investigation undertakingin an ideal way (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016). On the off chancethat it is found that the particular application does not give allperception ideas that are important for a satisfactory visualexamination of the information, one can come back to the past stage,refine the model and coordinate the improved method into state of artvisualization tool.

3DVisualization Service

Inview of the visualization work procedures that were explained in theprevious stage, in the last phase of the preparing model a 3Dvisualization administration is produced that computerizesfurthermore, institutionalizes the representation procedure. In thisway, the representation benefit framework that was displayed in theprevious section can be utilized (Laha, Sensharma, Schiffbauer, &ampBowman, 2012). The application contrast must be stretched out by thenew examination and representation strategies that were produced forthe particular application case. Moreover, structures for videosuccessions and question motion pictures must be planned that speakto the broke down datasets. Whenever vital, new investigation demandsfor the new errands must be created and to be incorporated into theframework (Velten,Wu, Jarabo, Masia, Barsi, Joshi, &amp Raskar, 2013). At long last,an assignment specific web interface must be created that givesinstinctive access to the administration functionality. Theadministration can be utilized by an expansive gathering of clients.At to begin with, restorative specialists also, scientists canutilize it to enhance and accelerate the investigation of themedicinal information. Since most strides of the examination andrepresentation process are computerized, the administration canlikewise be worked by other restorative staff who has less learningabout the restorative points of interest (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016).

Theadministration functionality can even be abused by specialists inlittler restorative offices who are not specialists for theparticular therapeutic undertaking. The administration assist permitsthe fuse of specialists from remote areas. There are twocircumstances which can require the arrival to the past advancementarrange. To begin with, it can be found that a perception workprocess is not suited for robotization (Bader, Kolb, Weaver, &ampOxman, 2016). At that point, the work process must be properlyadjusted and assessed with the comparing intelligent representationinstrument. The other circumstance can happen amid the consistent useof the administration when the predefined work process does not givea satisfactory representation for the present volume visualizationdataset (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016). For thissituation, a therapeutic physician ought to furthermore inspect theinformation with the intelligent 3D representation instrument.Subsequently, the perception benefit does not supplant the relatedintelligent representation application, yet, the two supplement eachother. The administration can be utilized for standard cases, and atthe point when issues happen, the intuitive application can berecommended (Bader, Kolb, Weaver, &amp Oxman, 2016).


Insummary, medical dataset visualization techniques have evolvedrapidly in the recent past. there are various tomographic techniquesthat have been used in the capturing of these medical data that areseemingly compatible with the current context of image dispositionand analysis. The role of technology cannot be underestimated heresince it has helped in the development of various facets of visualanalysis and interpretation (Bader, Kolb, Weaver, &amp Oxman, 2016).The techniques such as SPECT, CT, MRI, FMRI are some of the vitaltechniques used for capturing the medical data for visualization. Allthese techniques are chosen on the account of their capabilities andthe condition of the patient at hand. For example, a physician woulduse a CT instead of MRI on a patient who has metallic implants inthem for the fact that electromagnetic forces might interfere withthe whole context of metallic alignment in the body.

Onthe account of visualization techniques, the 3D visualization andvolume visualization goes hand in hand and cannot be separated asthey try to give the same type of information required for themedical analysis (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016). Medicalvisualization pipeline has been discussed in the research with keenattention to the notion of high level of investigations into how bestthe both volume and 3D visualization could be done. The pipelinestarts with the notion of building of an algorithm that allows forthe formation of basis of programs which are essentially important increating the visualization framework (Bader, Kolb, Weaver, &ampOxman, 2016). The next step is the notion of creation f applicationthat can allow for proper visualization and analysis of the wholecontext of medical volume dataset. The application can be built tosuit s specific kind of disease or can be modified to suite variousanalysis in the medical field. The problem-specific applications areusually based on the fact that there is need for specialization andconcentration into one particular issue on the body for example theconcentration on the cancer disease. The last step on the pipeline isthe automation of the application to allow for efficiency of theworkflows. Similarly, the automation process can be based on avariety of functionality of can be problem specific (Pycinski,Czajkowska, Badura, Juszczyk, &amp Pietka, 2016).

Thevisualization stage entails the conversion of the reprocessed 3Dvolume data to the 2D image (Shin,Orton, Collins, Doran, &amp Leach, 2013).This process is duly named volume visualization and hence consists ofthe mapping of the whole data volume to be presented into arenderable state where it is easy to create a 2D projection owing toits presentation. There are three distinct ways in which the volumedata could be visualized. These ways differ a great deal in terms oftheir mapping and also their rendering steps (Shneiderman,Plaisant, &amp Hesse, 2013). Thesesetups are put away and are added to the created surface work when arelating vertex arrangement happens. The position of the trianglevertices along the cell edges is processed by direct interjection.The walking solid shapes calculation can likewise be connected tofragmented volume information however then purported stair-casecurios happen because of the lofty move between voxels inside andoutside a divided locale.

Tosmother these antiques the portioned volume dataset and additionallythe produced work can be smoothed. A produced triangle work can bestraightforwardly rendered with traditional illustrations equipmentbolster. To accomplish a three-dimensional impression of thestructure the surface must be furthermore lit up and shaded.Therefore, the surface-based volume rendering method is frequentlyalluded to as surface shaded show (Tomar,&amp Agarwal, 2013).The benefit of surface-based rendering is its immediate support bythe 3D representation equipment. Be that as it may, consequently theera of a surface mesh is a costly pre-handling step. Besides, asurface model displays stand out certain structure of a dataset andnot the dataset in general. At the point when diverse structures of adataset ought to be examined all the while, a few surface crosssections must be produced and rendered in mix.

Itis important to understand the best ways in which medicalpractitioners are able to diagnose their patients in the mostaccurate way and give the best advice on the medication. Themilestone seen in the technological advancements in the graphics and3D visualization is applied in the context of medical sector with themain aim of knowing their immense significance (Cavalcanti,Rocha, &amp Vannier, 2014).The research includes the investigations of various researches acrossthe world with a view to bringing out the differences in thesetechniques and their importance to the medical field. More of thisresearch gives a detailed analysis of the tomographic imaging methodsin the medical field, the pipeline for medical visualization, therendering that is accelerated by the hardware and the multi-volumerendering which may be direct of flexible (Cevidanes,Bailey, Tucker, Styner, Phillips, &amp Turvey, 2014).Since these two techniques of visualization goes hand in hand, theirmedical field use is explored concurrently. The role of technology isvery immense in ensuring that the visualization techniques are welladdressed. Their development is based on the combination of both themedical knowledge and the technological advancement seen in thiscontext.


Abrahamsson,S., Chen, J., Hajj, B., Stallinga, S., Katsov, A. Y., Wisniewski, J.,… &amp Darzacq, X. (2013). Fast multicolor 3D imaging usingaberration-corrected multifocus microscopy. Naturemethods,10(1),60-63.

Amunts,K., Lepage, C., Borgeat, L., Mohlberg, H., Dickscheid, T., Rousseau,M. É., … &amp Shah, N. J. (2013). BigBrain: anultrahigh-resolution 3D human brain model. Science,340(6139),1472-1475.

Araki,K., Maki, K., Seki, K., Sakamaki, K., Harata, Y., Sakaino, R., … &ampSeo, K. (2014). Characteristics of a newly developeddentomaxillofacial X-ray cone beam CT scanner (CB MercuRay™):system configuration and physical properties. DentomaxillofacialRadiology.

Bader,C., Kolb, D., Weaver, J. C., &amp Oxman, N. (2016). Data-DrivenMaterial Modeling with Functional Advection for 3D Printing ofMaterially Heterogeneous Objects. 3DPrinting and Additive Manufacturing,3(2),71-79.

Berg,S., Ott, H., Klapp, S. A., Schwing, A., Neiteler, R., Brussee, N.,… &amp Kersten, M. (2013). Real-time 3D imaging of Haines jumps inporous media flow. Proceedingsof the National Academy of Sciences,110(10),3755-3759.

Binnewijzend,M. A., Kuijer, J. P., Benedictus, M. R., van der Flier, W. M., Wink,A. M., Wattjes, M. P., … &amp Barkhof, F. (2013). Cerebral bloodflow measured with 3D pseudocontinuous arterial spin-labeling MRimaging in Alzheimer disease and mild cognitive impairment: a markerfor disease severity. Radiology,267(1),221-230.

Borgo,R., Kehrer, J., Chung, D. H., Maguire, E., Laramee, R. S., Hauser,H., … &amp Chen, M. (2013). Glyph-based visualization:Foundations, design guidelines, techniques and applications.EurographicsState of the Art Reports,39-63.

Brodlie,K. W., Carpenter, L. A., Earnshaw, R. A., Gallop, J. R., Hubbold, R.J., Mumford, A. M., … &amp Quarendon, P. (Eds.). (2012).Scientificvisualization: techniques and applications.Springer Science &amp Business Media.

Brodlie,K., Osorio, R. A., &amp Lopes, A. (2012). A review of uncertainty indata visualization. In Expandingthe frontiers of visual analytics and visualization(pp. 81-109). Springer London.

Carrasco-Zevallos,O. M., Keller, B., Viehland, C., Shen, L., Waterman, G., Todorich,B., … &amp Toth, C. A. (2016). Live volumetric (4D) visualizationand guidance of in vivo human ophthalmic surgery with intraoperativeoptical coherence tomography. ScientificReports,6.

Cavalcanti,M. G. P., Rocha, S. S., &amp Vannier, M. W. (2014). Craniofacialmeasurements based on 3D-CT volume rendering: implications forclinical applications. DentomaxillofacialRadiology.

Cevidanes,L. H., Bailey, L. J., Tucker Jr, G. R., Styner, M. A., Mol, A.,Phillips, C. L., … &amp Turvey, T. (2014). Superimposition of 3Dcone-beam CT models of orthognathic surgery patients.DentomaxillofacialRadiology.

Chan,S., Conti, F., Salisbury, K., &amp Blevins, N. H. (2013). Virtualreality simulation in neurosurgery: technologies and evolution.Neurosurgery,72,A154-A164.

Chen,K. Y., Lyons, K., White, S., &amp Patel, S. (2013, October). uTrack:3D input using two magnetic sensors. In Proceedingsof the 26th annual ACM symposium on User interface software andtechnology(pp. 237-244). ACM.

Chen,M., Kaufman, A. E., &amp Yagel, R. (Eds.). (2012). Volumegraphics.Springer Science &amp Business Media.

Chiw,C., Kindlmann, G., Reppy, J., Samuels, L., &amp Seltzer, N. (2012,June). Diderot: a parallel DSL for image analysis and visualization.In Acmsigplan notices(Vol. 47, No. 6, pp. 111-120). ACM.

Crum,W. R., Hartkens, T., &amp Hill, D. L. G. (2014). Non-rigid imageregistration: theory and practice. TheBritish Journal of Radiology.

Dong,P., Haupert, S., Hesse, B., Langer, M., Gouttenoire, P. J., Bousson,V., &amp Peyrin, F. (2014). 3D osteocyte lacunar morphometricproperties and distributions in human femoral cortical bone usingsynchrotron radiation micro-CT images. Bone,60,172-185.

Donner,R., Menze, B. H., Bischof, H., &amp Langs, G. (2013). Globallocalization of 3D anatomical structures by pre-filtered HoughForests and discrete optimization. Medicalimage analysis,17(8),1304-1314.

Draenert,F. G., Coppenrath, E., Herzog, P., Müller, S., &amp Mueller-Lisse,U. G. (2014). Beam hardening artefacts occur in dental implant scanswith the NewTom® cone beam CT but not with the dental 4-rowmultidetector CT. DentomaxillofacialRadiology.

Earnshaw,R., &amp Wiseman, N. (2012). Anintroductory guide to scientific visualization.Springer Science &amp Business Media.

Ebner,M., Marone, F., Stampanoni, M., &amp Wood, V. (2013). Visualizationand quantification of electrochemical and mechanical degradation inLi ion batteries. Science,342(6159),716-720.

Eklund,A., Dufort, P., Forsberg, D., &amp LaConte, S. M. (2013). Medicalimage processing on the GPU–Past, present and future. Medicalimage analysis,17(8),1073-1094.

Evans,A., Romeo, M., Bahrehmand, A., Agenjo, J., &amp Blat, J. (2014). 3Dgraphics on the web: A survey. Computers&amp Graphics,41,43-61.

Fang,R., Pouyanfar, S., Yang, Y., Chen, S. C., &amp Iyengar, S. S.(2016). Computational health informatics in the big data age: asurvey. ACMComputing Surveys (CSUR),49(1),12.

Fedorov,A., Beichel, R., Kalpathy-Cramer, J., Finet, J., Fillion-Robin, J.C., Pujol, S., … &amp Buatti, J. (2012). 3D Slicer as an imagecomputing platform for the Quantitative Imaging Network. Magneticresonance imaging,30(9),1323-1341.

Ferroli,P., Tringali, G., Acerbi, F., Schiariti, M., Broggi, M., Aquino, D.,&amp Broggi, G. (2013). Advanced 3-dimensional planning inneurosurgery. Neurosurgery,72,A54-A62.

Fiolka,R., Shao, L., Rego, E. H., Davidson, M. W., &amp Gustafsson, M. G.(2012). Time-lapse two-color 3D imaging of live cells with doubledresolution using structured illumination. Proceedingsof the National Academy of Sciences,109(14),5311-5315.

Fong,E. L. S., Lamhamedi-Cherradi, S. E., Burdett, E., Ramamoorthy, V.,Lazar, A. J., Kasper, F. K., … &amp Amin, H. M. (2013). ModelingEwing sarcoma tumors in vitro with 3D scaffolds. Proceedingsof the National Academy of Sciences,110(16),6500-6505.

Fout,N., &amp Ma, K. L. (2012). An adaptive prediction-based approach tolossless compression of floating-point volume data. IEEETransactions on Visualization and Computer Graphics,18(12),2295-2304.

Gao,G., Penney, G., Ma, Y., Gogin, N., Cathier, P., Arujuna, A., … &ampHancock, J. (2012). Registration of 3D trans-esophagealechocardiography to X-ray fluoroscopy using image-based probetracking. Medicalimage analysis,16(1),38-49.

Garyfallidis,E., Brett, M., Amirbekian, B., Rokem, A., Van Der Walt, S.,Descoteaux, M., &amp Nimmo-Smith, I. (2014). Dipy, a library for theanalysis of diffusion MRI data. Frontiersin neuroinformatics,8,8.

Garyfallidis,E., Brett, M., Amirbekian, B., Rokem, A., Van Der Walt, S.,Descoteaux, M., &amp Nimmo-Smith, I. (2014). Dipy, a library for theanalysis of diffusion MRI data. Frontiersin neuroinformatics,8,8.

Gee,A., Prager, R., Treece, G., Cash, C., &amp Berman, L. (2014).Processing and visualizing three-dimensional ultrasound data. TheBritish journal of radiology.

Gouws,A., Woods, W., Millman, R., Morland, A., &amp Green, G. (2015).DataViewer3D: an open-source, cross-platform multi-modal neuroimagingdata visualization tool. Pythonin Neuroscience,86.

Gross,B. C., Erkal, J. L., Lockwood, S. Y., Chen, C., &amp Spence, D. M.(2014). Evaluation of 3D printing and its potential impact onbiotechnology and the chemical sciences. Analyticalchemistry,86(7),3240-3253.

Grottel,S., Krone, M., Müller, C., Reina, G., &amp Ertl, T. (2015).MegaMol—A Prototyping Framework for Particle-Based Visualization.IEEEtransactions on visualization and computer graphics,21(2),201-214.

Heiland,M., Schulze, D., Adam, G., &amp Schmelzle, R. (2014). 3D-imaging ofthe facial skeleton with an isocentric mobile C-arm system (SiremobilIso-C3D). DentomaxillofacialRadiology.

Hirsch,E., Wolf, U., Heinicke, F., &amp Silva, M. A. G. (2014). Dosimetryof the cone beam computed tomography Veraviewepocs 3D compared withthe 3D Accuitomo in different fields of view. DentomaxillofacialRadiology.

Hirsch,S., Guo, J., Reiter, R., Papazoglou, S., Kroencke, T., Braun, J., &ampSack, I. (2014). MR Elastography of the Liver and the Spleen Using aPiezoelectric Driver, Single‐ShotWave‐FieldAcquisition, and Multifrequency Dual Parameter Reconstruction.Magneticresonance in medicine,71(1),267-277.

Hirsch,S., Guo, J., Reiter, R., Papazoglou, S., Kroencke, T., Braun, J., &ampSack, I. (2014). MR Elastography of the Liver and the Spleen Using aPiezoelectric Driver, Single‐ShotWave‐FieldAcquisition, and Multifrequency Dual Parameter Reconstruction.Magneticresonance in medicine,71(1),267-277.

Höhne,K. H., Fuchs, H., &amp Pizer, S. M. (Eds.). (2012). 3Dimaging in medicine: algorithms, systems, applications(Vol. 60). Springer Science &amp Business Media.

Holzinger,A., Dehmer, M., &amp Jurisica, I. (2014). Knowledge discovery andinteractive data mining in bioinformatics-state-of-the-art, futurechallenges and research directions. BMCbioinformatics,15(6),1.

Hopp,T., Baltzer, P., Dietzel, M., Kaiser, W. A., &amp Ruiter, N. V.(2012). 2D/3D image fusion of X-ray mammograms with breast MRI:visualizing dynamic contrast enhancement in mammograms. Internationaljournal of computer assisted radiology and surgery,7(3),339-348.

Hornung,A., Wurm, K. M., Bennewitz, M., Stachniss, C., &amp Burgard, W.(2013). OctoMap: An efficient probabilistic 3D mapping frameworkbased on octrees. AutonomousRobots,34(3),189-206.

Huang,F., Sirinakis, G., Allgeyer, E. S., Schroeder, L. K., Duim, W. C.,Kromann, E. B., … &amp Lessard, M. (2016). Ultra-high resolution3D imaging of whole cells. Cell,166(4),1028-1040.

Idiyatullin,D., Corum, C. A., Nixdorf, D. R., &amp Garwood, M. (2014). Intraoralapproach for imaging teeth using the transverse B1 field componentsof an occlusally oriented loop coil. Magneticresonance in medicine,72(1),160-165.

Ieiri,S., Uemura, M., Konishi, K., Souzaki, R., Nagao, Y., Tsutsumi, N.,… &amp Tanoue, K. (2012). Augmented reality navigation system forlaparoscopic splenectomy in children based on preoperative CT imageusing optical tracking device. Pediatricsurgery international,28(4),341-346.

James,A. P., &amp Dasarathy, B. V. (2014). Medical image fusion: A surveyof the state of the art. InformationFusion,19,4-19.

Johnson,K. M., Fain, S. B., Schiebler, M. L., &amp Nagle, S. (2013).Optimized 3D ultrashort echo time pulmonary MRI. Magneticresonance in medicine,70(5),1241-1250.

Johnson,K. M., Fain, S. B., Schiebler, M. L., &amp Nagle, S. (2013).Optimized 3D ultrashort echo time pulmonary MRI. Magneticresonance in medicine,70(5),1241-1250.

Kehrer,J., &amp Hauser, H. (2013). Visualization and visual analysis ofmultifaceted scientific data: A survey. IEEEtransactions on visualization and computer graphics,19(3),495-513.

Kelm,B. M., Wels, M., Zhou, S. K., Seifert, S., Suehling, M., Zheng, Y., &ampComaniciu, D. (2013). Spine detection in CT and MR using iteratedmarginal space learning. Medicalimage analysis,17(8),1283-1292.

Kersten-Oertel,M., Jannin, P., &amp Collins, D. L. (2013). The state of the art ofvisualization in mixed reality image guided surgery. ComputerizedMedical Imaging and Graphics,37(2),98-112.

Kessler,M. L. (2014). Image registration and data fusion in radiationtherapy. TheBritish journal of radiology.

Kim,M. H., Rushmeier, H., Ffrench, J., Passeri, I., &amp Tidmarsh, D.(2014). Hyper3D: 3D graphics software for examining culturalartifacts. Journalon Computing and Cultural Heritage (JOCCH),7(3),14.

Kim,S. H., &amp Chung, K. Y. (2015). Medical information service systembased on human 3D anatomical model. MultimediaTools and Applications,74(20),8939-8950.

Kruger,R. A., Kuzmiak, C. M., Lam, R. B., Reinecke, D. R., Del Rio, S. P., &ampSteed, D. (2013). Dedicated 3D photoacoustic breast imaging. Medicalphysics,40(11),113301.

Kumar,V., Ludlow, J. B., Mol, A., &amp Cevidanes, L. (2014). Comparison ofconventional and cone beam CT synthesized cephalograms.DentomaxillofacialRadiology.

Laha,B., Sensharma, K., Schiffbauer, J. D., &amp Bowman, D. A. (2012).Effects of immersion on visual analysis of volume data. IEEETransactions on Visualization and Computer Graphics,18(4),597-606.

Limaye,A. (2012, October). Drishti-volume exploration and presentation tool.In ProcSpie(Vol. 8506, pp. 85060X-85060X).

Lin,X., Wang, J., Han, F., Fu, J., &amp Li, A. (2012). Analysis ofeighty-one cases with breast lesions using automated breast volumescanner and comparison with handheld ultrasound. Europeanjournal of radiology,81(5),873-878.

Lüsebrink,F., Wollrab, A., &amp Speck, O. (2013). Cortical thicknessdetermination of the human brain using high resolution 3T and 7T MRIdata. Neuroimage,70,122-131.

Maier-Hein,L., Mountney, P., Bartoli, A., Elhawary, H., Elson, D., Groch, A.,… &amp Stoyanov, D. (2013). Optical techniques for 3D surfacereconstruction in computer-assisted laparoscopic surgery. Medicalimage analysis,17(8),974-996.

Mailly,P., Aliane, V., Groenewegen, H. J., Haber, S. N., &amp Deniau, J. M.(2013). The rat prefrontostriatal system analyzed in 3D: evidence formultiple interacting functional units. TheJournal of neuroscience,33(13),5718-5727.

Marcellin,M. W., Bilgin, A., Lalgudi, H. G., &amp Nadar, M. S. (2012). U.S.Patent No. 8,126,279.Washington, DC: U.S. Patent and Trademark Office.

Marcus,D. S., Harms, M. P., Snyder, A. Z., Jenkinson, M., Wilson, J. A.,Glasser, M. F., … &amp Hodge, M. (2013). Human Connectome Projectinformatics: quality control, database services, and datavisualization. Neuroimage,80,202-219.

Markelj,P., Tomaževič, D., Likar, B., &amp Pernuš, F. (2012). A review of3D/2D registration methods for image-guided interventions. Medicalimage analysis,16(3),642-661.

McConnell,G., Trägårdh, J., Amor, R., Dempster, J., Reid, E., &amp Amos, W.B. (2016). A novel optical microscope for imaging large embryos andtissue volumes with sub-cellular resolution throughout. eLife,5,e18659.

Meijering,E. (2012). Cell segmentation: 50 years down the road [life sciences].IEEESignal Processing Magazine,29(5),140-145.

Mischkowski,R. A., Scherer, P., Ritter, L., Neugebauer, J., Keeve, E., &ampZöller, J. E. (2014). Diagnostic quality of multiplanar reformationsobtained with a newly developed cone beam device for maxillofacialimaging. DentomaxillofacialRadiology.

Morgeneyer,T. F., Helfen, L., Mubarak, H., &amp Hild, F. (2013). 3D digitalvolume correlation of synchrotron radiation laminography images ofductile crack initiation: an initial feasibility study. ExperimentalMechanics,53(4),543-556.

Mori,S., Endo, M., Nishizawa, K., Murase, K., Fujiwara, H., &amp Tanada,S. (2014). Comparison of patient doses in 256-slice CT and 16-sliceCT scanners. TheBritish journal of radiology.

Morrison,R. J., Hollister, S. J., Niedner, M. F., Mahani, M. G., Park, A. H.,Mehta, D. K., … &amp Green, G. E. (2015). Mitigation oftracheobronchomalacia with 3D-printed personalized medical devices inpediatric patients. Sciencetranslational medicine,7(285),285ra64-285ra64.

Nelson,S. J., Kurhanewicz, J., Vigneron, D. B., Larson, P. E., Harzstark, A.L., Ferrone, M., … &amp Reed, G. (2013). Metabolic imaging ofpatients with prostate cancer using hyperpolarized [1-13C] pyruvate.Sciencetranslational medicine,5(198),198ra108-198ra108.

Nkenke,E., Zachow, S., Benz, M., Maier, T., Veit, K., Kramer, M., … &ampLell, M. (2014). Fusion of computed tomography data and optical 3Dimages of the dentition for streak artefact correction in thesimulation of orthognathic surgery. DentomaxillofacialRadiology.

Oliveira,F. P., &amp Tavares, J. M. R. (2014). Medical image registration: areview. Computermethods in biomechanics and biomedical engineering,17(2),73-93.

Peng,H., Bria, A., Zhou, Z., Iannello, G., &amp Long, F. (2014).Extensible visualization and analysis for multidimensional imagesusing Vaa3D. Natureprotocols,9(1),193-208.

Peng,H., Tang, J., Xiao, H., Bria, A., Zhou, J., Butler, V., … &ampMitra, A. (2014). Virtual finger boosts three-dimensional imaging andmicrosurgery as well as terabyte volume image visualization andanalysis. Naturecommunications,5.

Peng,H., Tang, J., Xiao, H., Bria, A., Zhou, J., Butler, V., … &ampMitra, A. (2014). Virtual finger boosts three-dimensional imaging andmicrosurgery as well as terabyte volume image visualization andanalysis. Naturecommunications,5.

Post,F. H., Nielson, G., &amp Bonneau, G. P. (Eds.). (2012). Datavisualization: the state of the art(Vol. 713). Springer Science &amp Business Media.

Potter,K., Rosen, P., &amp Johnson, C. R. (2012). From quantification tovisualization: A taxonomy of uncertainty visualization approaches. InUncertaintyQuantification in Scientific Computing(pp. 226-249). Springer Berlin Heidelberg.

Preim,B., &amp Botha, C. P. (2013). VisualComputing for Medicine: Theory, Algorithms, and Applications.Newnes.

Provost,J., Papadacci, C., Arango, J. E., Imbault, M., Fink, M., Gennisson,J. L., … &amp Pernot, M. (2014). 3D ultrafast ultrasound imagingin vivo. Physicsin medicine and biology,59(19),L1.

Pycinski,B., Czajkowska, J., Badura, P., Juszczyk, J., &amp Pietka, E.(2016). Time-Of-Flight Camera, Optical Tracker and ComputedTomography in Pairwise Data Registration. PloSone,11(7),e0159493.

Raghupathi,W., &amp Raghupathi, V. (2014). Big data analytics in healthcare:promise and potential. HealthInformation Science and Systems,2(1),1.

Rodt,T., Bartling, S. O., Zajaczek, J. E., Vafa, M. A., Kapapa, T.,Majdani, O., … &amp Kaminsky, J. (2014). Evaluation of surface andvolume rendering in 3D-CT of facial fractures. DentomaxillofacialRadiology.

Saalfeld,P., Stojnic, A., Preim, B., &amp Oeltze, S. (2016). „Semi-Immersive3D Sketching of Vascular Structures for Medical Education “. InProceedingsof Euroraphics Workshop on Visual Computing for Biology and Medicine.

Schäfer,A., Forstmann, B. U., Neumann, J., Wharton, S., Mietke, A., Bowtell,R., &amp Turner, R. (2012). Direct visualization of the subthalamicnucleus and its iron distribution using high‐resolutionsusceptibility mapping. Humanbrain mapping,33(12),2831-2842.

Seeram,E. (2015). Computedtomography: physical principles, clinical applications, and qualitycontrol.Elsevier Health Sciences.

Seeram,E. (2015). Computedtomography: physical principles, clinical applications, and qualitycontrol.Elsevier Health Sciences.

Shin,H. C., Orton, M. R., Collins, D. J., Doran, S. J., &amp Leach, M. O.(2013). Stacked autoencoders for unsupervised feature learning andmultiple organ detection in a pilot study using 4D patient data. IEEEtransactions on pattern analysis and machine intelligence,35(8),1930-1943.

Shneiderman,B., Plaisant, C., &amp Hesse, B. W. (2013). Improving healthcarewith interactive visualization. Computer,46(5),58-66.

Spottiswoode,B. S., Van den Heever, D. J., Chang, Y., Engelhardt, S., Du Plessis,S., Nicolls, F., … &amp Gretschel, A. (2013). Preoperativethree-dimensional model creation of magnetic resonance brain imagesas a tool to assist neurosurgical planning. Stereotacticand functional neurosurgery,91(3),162-169.

Spottiswoode,B. S., Van den Heever, D. J., Chang, Y., Engelhardt, S., Du Plessis,S., Nicolls, F., … &amp Gretschel, A. (2013). Preoperativethree-dimensional model creation of magnetic resonance brain imagesas a tool to assist neurosurgical planning. Stereotacticand functional neurosurgery,91(3),162-169.

Stankovic,Z., Allen, B. D., Garcia, J., Jarvis, K. B., &amp Markl, M. (2014).4D flow imaging with MRI. Cardiovasculardiagnosis and therapy,4(2),173-192.

Tan,T. C., Zeng, X., Jiao, Y., Wang, L., Wei, Q., Thiele, K., … &ampHung, J. (2016). Three-dimensional field optimization method:clinical validation of a novel color Doppler method for quantifyingmitral regurgitation. Journalof the American Society of Echocardiography,29(10),926-934.

Thorwarth,D., Müller, A. C., Pfannenberg, C., &amp Beyer, T. (2013). CombinedPET/MR imaging using 68Ga-DOTATOC for radiotherapy treatment planningin meningioma patients. In Theranostics,Gallium-68, and Other Radionuclides(pp. 425-439). Springer Berlin Heidelberg.

Tian,L., &amp Waller, L. (2015). 3D intensity and phase imaging fromlight field measurements in an LED array microscope. Optica,2(2),104-111.

Tobon-Gomez,C., De Craene, M., Mcleod, K., Tautz, L., Shi, W., Hennemuth, A., …&amp Lutz, A. (2013). Benchmarking framework for myocardial trackingand deformation algorithms: An open access database. Medicalimage analysis,17(6),632-648.

Tomar,D., &amp Agarwal, S. (2013). A survey on Data Mining approaches forHealthcare. InternationalJournal of Bio-Science and Bio-Technology,5(5),241-266.

Varallyay,C. G., Nesbit, E., Fu, R., Gahramanov, S., Moloney, B., Earl, E., …&amp Neuwelt, E. A. (2013). High-resolution steady-state cerebralblood volume maps in patients with central nervous system neoplasmsusing ferumoxytol, a superparamagnetic iron oxide nanoparticle.Journalof Cerebral Blood Flow &amp Metabolism,33(5),780-786.

Velten,A., Wu, D., Jarabo, A., Masia, B., Barsi, C., Joshi, C., … &ampRaskar, R. (2013). Femto-photography: capturing and visualizing thepropagation of light. ACMTransactions on Graphics (TOG),32(4),44.

Volonté,F., Pugin, F., Buchs, N. C., Spaltenstein, J., Hagen, M., Ratib, O.,&amp Morel, P. (2013). Console-integrated stereoscopic OsiriX 3Dvolume-rendered images for da Vinci colorectal robotic surgery.Surgicalinnovation,20(2),158-163.

Wang,J., Chen-Wiegart, Y. C. K., Eng, C., Shen, Q., &amp Wang, J. (2016).Visualization of anisotropic-isotropic phase transformation dynamicsin battery electrode particles. NatureCommunications,7.

Ware,C. (2012). Informationvisualization: perception for design.Elsevier.

Würslin,C., Schmidt, H., Martirosian, P., Brendle, C., Boss, A., Schwenzer,N. F., &amp Stegger, L. (2013). Respiratory motion correction inoncologic PET using T1-weighted MR imaging on a simultaneouswhole-body PET/MR system. Journalof Nuclear Medicine,54(3),464-471.

Xia,M., Wang, J., &amp He, Y. (2013). BrainNet Viewer: a networkvisualization tool for human brain connectomics. PloSone,8(7),e68910.

Yu,L., Efstathiou, K., Isenberg, P., &amp Isenberg, T. (2012).Efficient structure-aware selection techniques for 3D point cloudvisualizations with 2DOF input. IEEETransactions on Visualization and Computer Graphics,18(12),2245-2254.

Zhao,F., &amp Xie, X. (2013). An overview of interactive medical imagesegmentation. Annalsof the BMVA,2013(7),1-22.