Palaeontologia Electronica · Palaeontologia Electronica palaeo-electronica.org ... (Egerton), a shark from the Late ... rare specimens and type materials in university and museum

Palaeontologia Electronica palaeo-electronica.org

A palaeobiologist’s guide to ‘virtual’ micro-CT preparation

Richard Leslie Abel, Carolina Rettondini Laurini, and Martha Richter

ABSTRACT

This paper provides a brief but comprehensive guide to creating, preparing anddissecting a ‘virtual’ fossil, using a worked example to demonstrate some standarddata processing techniques. Computed tomography (CT) is a 3D imaging modality forproducing ‘virtual’ models of an object on a computer. In the last decade, CT technol-ogy has greatly improved, allowing bigger and denser objects to be scanned increas-ingly rapidly. The technique has now reached a stage where systems can facilitatelarge-scale, non-destructive comparative studies of extinct fossils and their living rela-tives. Consequently the main limiting factor in CT-based analyses is no longer scan-ning, but the hurdles of data processing (see disclaimer). The latter comprises thetechniques required to convert a 3D CT volume (stack of digital slices) into a virtualimage of the fossil that can be prepared (separated) from the matrix and ‘dissected’into its anatomical parts. This technique can be applied to specimens or part of speci-mens embedded in the rock matrix that until now have been otherwise impossible tovisualise. This paper presents a suggested workflow explaining the steps required,using as example a fossil tooth of Sphenacanthus hybodoides (Egerton), a shark fromthe Late Carboniferous of England. The original NHMUK copyrighted CT slice stackcan be downloaded for practice of the described techniques, which include segmenta-tion, rendering, movie animation, stereo-anaglyphy, data storage and dissemination.Fragile, rare specimens and type materials in university and museum collections cantherefore be virtually processed for a variety of purposes, including virtual loans, web-site illustrations, publications and digital collections. Micro-CT and other 3D imagingtechniques are increasingly utilized to facilitate data sharing among scientists and oneducation and outreach projects. Hence there is the potential to usher in a new era ofglobal scientific collaboration and public communication using specimens in museumcollections.

Richard Leslie Abel, MSK Laboratory, Department of Surgery and Cancer, Charring Cross Hospital, Imperial College, W6 8RF London, United Kingdom. [email protected] and Analysis Centre, Mineralogy Department, Natural History Museum, Cromwell Road, SW7 5BD London, United Kingdom. Carolina Rettondini Laurini, Laboratório de Paleontologia Departamento de Biologia FFCLRP - USP, Av. Bandeirantes, 3900 - CEP 14040-901, Bairro Monte Alegre, Ribeirão Preto, SP, Brazil. [email protected]

PE Article Number: 15.2.6TCopyright: Palaeontological Association May 2012Submission: 12 May 2011. Acceptance: 25 April 2012

Abel, Richard Leslie, Laurini, Carolina Rettondini, and Richter, Martha 2012. A palaeobiologist’s guide to ‘virtual’ micro-CT preparation. Palaeontologia Electronica Vol. 15, Issue 2;6T,17p; palaeo-electronica.org/content/issue-2-2012-technical-articles/233-micro-ct-workflow

ABEL ET AL.: MICRO-CT WORKFLOW

Martha Richter, Department of Palaeontology, Natural History Museum, Cromwell Road, SW7 5BD London, United Kingdom. [email protected]

Keywords: Computed tomography; micro-CT scan; X-ray microtomography; fossil shark tooth; segmenta-tion and virtual preparation

INTRODUCTION: WHAT ISCOMPUTED TOMOGRAPHY?

Computed tomography (CT) is a non-destruc-tive radiographic imaging technique for producing3D computerised models of an object. The firstsystem was developed by Godfrey Hounsfield(Hounsfield, 1973) and Allan McCormack (McCor-mack, 1963) and went online in 1972 at the Atkin-son Morley's Hospital (London, UK). During the last40 years, CT scanners have undergone severalredesigns - often referred to as generations - butthe basic concept has remained the same. All CTscanners reconstruct digital cross sections (slices)of an object, which can be stacked to create 3Dvolumes rather like traditional serial sectioningtechniques (Sollas, 1904; Croft, 1950); hence‘tomography,’ from the Greek tomos (slice) andgraphein (to write). The resulting 3D volumes canbe used to create ‘virtual’ computerised images ofspecimens that can be manipulated, sectioned,prepared, dissected and measured as though inthe hand, but – unlike handheld specimens – withinternal as well as external morphology. Thisallows access to the morphological informationcontained inside fragile, rare, valuable or small

specimens, including both extinct fossils and extantcomparative material. However, creating, manipu-lating and preparing a good quality virtual fossil ona computer can be very time-consuming, and diffi-cult specimens can take 40 hours to process. Thisprocess greatly exceeds the time required for col-lecting the original CT volume, which varies from15-120 minutes (depending on the system). Hencedata processing is usually the main limiting factorin CT-based research, as opposed to scanning. Agood result also depends on certain qualities of thematerial to be scanned (see below).

FOSSIL MATERIAL UTILIZED

A single fossil shark tooth of Sphenacanthushybodoides (Egerton, 1853) (accession numberNHMUK PV P.1322) was utilized to demonstratethe potential benefits of micro-CT investigation.This freshwater shark species was found at Long-ton, Staffordshire, England (see Maisey, 1982;Dick, 1998) and lived during the Late Carbonifer-ous approximately 315 MYA. Half of the fossil isembedded in matrix, and its digital preparationrequires the full repertoire of extraction procedures(Figure 1.1). Yet the relatively simple structure of

FIGURE 1. Micro-CT can be used to” virtually” extract fossils. (1.1) Photo of the shark tooth in matrix and (1.2) sec-tioned micro- CT rendering after virtual preparation from rock matrix.

2

PALAEO-ELECTRONICA.ORG

the tooth allows a virtual preparation to be carriedout reasonably quickly, approximately 16-40 hoursfor a beginner or 4 hours for an expert (Figure 1.2).The original NHMUK (London) copyrighted micro-CT dataset of the fossil, comprising a tiff stack, canbe be obtained from Martha Richter for private(non-commercial) practice of the described tech-niques. In order to use the data to practice, onemust first understand how CT data is collected andwhat it actually consists of.

HOW DOES A CT SCANNER WORK?

All micro-CT systems employ the same basicprinciples. Check the glossary for definition ofterms. First the fossil is radiographed from manydifferent directions to map density distribution, asmeasured by X-ray transmission. Then computersoftware is used to reconstruct the 2D radiographsinto a 3D volume. During this step the data is usu-ally processed to remove artefacts (mistakes) thatinvariably appear in CT scans, e.g., ‘noise.’

Radiography

The Sphenacanthus tooth was scanned usingan HMX-ST 225 CT System (Nikon Metrology,Tring, UK). The system consists of three key com-ponents: an X-ray source, a turntable and an X-raydetector panel (2000x2000 pixels). The X-raysource produces a cone-shaped X-ray beam bybombarding a metal target with electrons, gener-ated by passing a high-energy electric currentthrough a tungsten filament. The cone is focusedonto a fossil, which is mounted on the turntableand rotated through 360º. During the rotation thedetector panel collects a series of two-dimensionalprojections (radiographs) usually at between 0.1ºto 0.05º intervals. The projections measure theamount of X-ray energy transmitted by the fossil. Inorder to produce a radiograph with high contrastand brightness, the X-ray beam must adequatelypenetrate the sample without over-exposing thepanel. Hence the energy of the X-rays (determinedby the voltage and current) has to be carefullyselected by the CT user. Penetration of the beam(and projection contrast) is largely determined bythe voltage, whilst the number of X-rays (and pro-jection brightness) is mainly determined by the cur-rent. The fossil shark tooth was scanned with an X-ray beam set at 180 kV and 138 µA. A larger ordenser fossil would require higher settings. It isworth noting that CT scans of dense specimens athigh x-ray energy are likely to include beam-hard-ening artefacts, which can be reduced using filtra-tion techniques.

Computer Reconstruction

After scanning, a computer is used to line upand centre the projections in a radial pattern, ratherlike the spokes of a bicycle wheel. Each row of pix-els in the detector panel will become a slice, so across-section of the specimen is created from eachline of pixels in the radiograph using a methodcalled back projection. Essentially, the projectionsfor each slice are converted to digital profiles thensmeared across each other to create digital imagesof the corresponding CT slices (see Figure 2).Each slice is made up of voxels, i.e., three-dimen-sional pixels. A CT reconstruction (or volume) isthus essentially a matrix (rather like a Rubik’scube). Each voxel is assigned a CT number or greyvalue. This value is based on the X-ray density(i.e., linear attenuation coefficient) of the materialsbeing scanned (Zonneveld, 1987). The coefficientis a measure of the absorption (and scatter) of anX-ray beam as it passes through a material and isdependent largely on the density (but also thick-ness and atomic number). Thus a CT scan is akinto a 3D extension of a greyscale digital photo-graph, but based on X-ray transmission rather thanvisible light. The size of the voxels is determined bythe magnification of the projected fossil image onthe panel. The turntable can be moved to varyingdistances between the X-ray source and detector

FIGURE 2. A simple example of back projection basedon two digital projections at 90o to each other. Corre-sponding rows of pixels from each projection are usedto create an X-ray transmission profile. Profiles areused to create digital images of the row that are thenback projected (i.e., smeared) onto one another to cre-ate a 3D CT slice. In this example the cylindrical objectsappear in the image as cubes, but the shape would beresolved if more projections were analysed. This pro-cess is repeated for every slice (pixel row) in the scan.

3


until the projection of the fossil fills the panel. Thebest magnification possible therefore is one wherethe voxel size is 1/2000th of the width or height ofthe fossil (whichever is greater). Voxel size is a keyconcept in CT, because the dimensions (largely)determine the spatial resolution of the scan, i.e.,the ability to resolve two objects of similar densitythat are situated close to one another. This value isusually between two and five times the size of thevoxels. For the shark tooth, voxel size was 0.009mm, and the effective spatial resolution was ≥0.050 mm.

Artefact Reduction

Micro-CT computer reconstruction is prone toartefacts (errors and mistakes). Artefacts are usu-ally manifested as a blurring of material boundariesor a grainy appearance. The errors are caused byseveral aspects of CT such as the matrix (voxel)structure of the volumes, the properties of X-raybeams, X-ray scatter and electronic errors in thedigital detector panel. The most commonly encoun-tered artefacts, partial volume averaging effectsand noise, greatly affect the spatial resolution. Vol-ume averaging results from numerous linear atten-uation coefficients (i.e., materials of differentdensity) occupying a single voxel, being repre-sented by an averaged grey value. This phenome-non leads to a gradient of CT values at materialinterfaces (e.g., fossil and matrix) that appear tothe naked eye as blurring of boundaries (Spoor etal., 1993; Zonneveld, 1987). The greater the sizeof the voxels relative to the width of anatomical fea-tures the greater the extent of the blurring. Finefeatures below the size of a voxel can become tooblurred to visualise, or even disappear completely(but see McColl et al., 2006).

At very small voxel size, noise often becomesthe factor limiting spatial resolution. All scans con-tain noise: unusually bright and dark voxels aresuperimposed over the CT data, which gives CTscans a grainy or speckled appearance, leading toblurring of material boundaries. X-ray scatter,which causes the photons to follow a non-linearpath to the detector, can cause speckling, as canvariation in the ability of pixels in the panel todetect X-rays. The extent of the speckling is deter-mined by the signal (i.e., X-ray energy) to noise(i.e., error) ratio.

Another artefact commonly associated withdense objects like fossils is beam hardening, whichis caused by variation in the energy of the X-raybeam. Micro-CT scanners produce polychromaticbeams, i.e., X-rays of more than one wavelength.

Hardening is the process of selective absorption oflow energy X-rays from the polychromatic beam.As low energy X-rays are absorbed or scattered bya fossil, the beam becomes progressively harder ormore penetrating. Thus the material at the edge ofa fossil appears to be more dense (i.e., greater X-ray absorption) than the centre (i.e., lower X-rayabsorption). The artefact is manifested in CT slicesas a cupping of the CT numbers: a ring of exces-sively bright voxels around the edge of the fossilreducing to excessively dark voxels at the centre(see Ronan et al., 2010, figure 2). It is possible toreduce beam-hardening artefacts by scanning per-pendicular to the long axis of the fossil (i.e., mountit standing up) or placing a copper filter in betweenthe X-ray source and the fossil. The filter removeslow energy X-rays and reduces the cupping arte-fact. By experimenting with different thicknessesand exposure times for the projections, it is usuallypossible to remove beam-hardening artefacts alto-gether. Alternatively, data processing algorithmscan be used to remove the artefact post-scanning.To do this, it is possible to plot a transect acrossthe centre of the specimen and calculate a secondor third order polynomial that describes the varia-tion (cupping) in grey values. The inverse of thispolynomial can then be applied to even out thegrey values. For a full discussion of beam-harden-ing artefacts and their removal see Ronan et al.(2010, 2011).

HOW AND WHERE DO I GET A CT SCAN?

The best option for obtaining a CT scan is toapproach a centre of excellence, which carries outcollaborative research with external users. Exam-ples in the UK include: Imaging and Analysis Cen-tre, Natural History Museum, London; Vis-µ,Southampton University; the School of Materials,University of Manchester; Department of Archaeol-ogy and Anthropology, University of Bristol; and CT@ SIMBIOS, University of Abertay. Collaborativescanning has many benefits because users canrely on experienced computed-tomographers toobtain high density-contrast scans, and then con-centrate on the downstream data processing,which usually requires special anatomical knowl-edge. Furthermore, this type of contract scanningis actually very cheap. Charges vary but users canexpect to pay between £10-120 per scan (see Abelet al., 2011). It is possible to include more than onespecimen in a single scan, providing the voxel sizewill be small enough to image the anatomical fea-tures of interest. Alternatively, systems are becom-ing more affordable for scientists, so research

4


groups can purchase their own as many Universi-ties and Museums are choosing to do. A bench topSkyScan system or a cabinet Nikon Metrology sys-tem will cost about £150-250K.

HOW DO I PREPARE AND DISSECT A ‘VIRTUAL’ FOSSIL?

Since scans are becoming increasingly easyand cheap to obtain, a lack of training in imagingprocessing appears to be the major factor limitingCT based research. Software manuals are usuallyvery technical and do not provide a workflow. Thiscontribution hopes to bridge that knowledge gap bypresenting a workflow that users can follow with anexemplar CT data set. Users have a wide variety ofsoftware packages, so this section contains a gen-eral description of the tools available to research-ers. Specific advice on suitable software choices isprovided later.

The example explored herein is a virtual prep-aration and dissection of the fossil shark tooth scandescribed above. The techniques suggested herecan be applied for any CT scan of a specimen,either extinct or extant, by following a systematicworkflow (Table 1). The tooth was partially embed-ded in rock matrix, and one of the objectives was toinvestigate the internal network of spaces originallyoccupied by vascular channels. Virtual preparationand dissection of the specimen - in this case, thewhole fossil tooth and its internal vascular canalsystem - involved separating a scan into regions of

interest (ROI). These corresponded to rock matrix,dental tissues like dentine and enameloid and, vas-cular canals (voids). The process of dividing thevoxels in a 3D volume between discrete ROIs istermed segmentation. Regions were identified bycommon properties such as voxel grey value and/or location. Specifically, segmentation of the fossilwas carried out in four steps using: contrastenhancement; surface determination; ROI growingand masking tools.

Density Contrast Enhancement

A contrast enhancement was applied to thewhole stack of X-ray slices in order to better dis-criminate the grey values that represented fossiland matrix. A frequency distribution plot of voxelgrey values (i.e., grey values vs. number of voxels)was stretched and compressed by altering thefloating-point range. This numbering system regis-ters grey values as decimals rather than integers inorder to economise on the size and speed of thereconstruction (see Cline et al., 1998; Xu andMueller, 2005). When the data is read into CT soft-ware the decimals are displayed as integers forease of visualisation and manipulation by the enduser. Reducing the upper end of the floating-pointrange (i.e., removing the brightest voxels)stretched the spread of mid-range grey values(e.g., fossil) at the cost of reducing the range ofhigher CT numbers (e.g., matrix). Producing afequency distribution plot (Figure 3) and slice stack

TABLE 1. Ten step workflow for virtual fossil preparation.

Task Step ProcessR

eco

nst

ruct

1 Reduce blurring by applying noise reduction (see method)

2 Reduce cupping by applying beam hardening correction (see method)

3 Increase contrast by stretching floating point range of fossil grey values

Ren

der

4 Segment fossil from matrix by applying global threshold (Fig.3A)

5 Segment fossil form non-adjacent matrix using region grower tool (Fig. 3B)

6 Remove adjacent matrix with a masking tool (Fig.3C)

7 Repeat steps 4-7 to segment anatomical features (Fig. 3D)

8 Apply lighting and false colour to create virtual specimen see (Fig. 4)

9 Animate rendered 3D model using a keyframer (Fig. 6)

Store 10 Archive slice stack as 8 Bit BMP, 16 Bit DICOM or 32 Bit vol files.

5


(Figure 4.1) with clearly distinguished matrix andfossil phases. This procedure greatly simplified andincreased the speed of the rest of the segmenta-tion process.

Surface Determination

The simplest and quickest way to segment anROI is to separate the maxima (nodes) in the greyvalue frequency distribution plot (Figure 3) by

applying a global threshold at the minima (valleys),which separate the nodes (See McColl et al., 2006;Reissis and Abel, 2012). Any voxels with a greyvalue above the threshold value are grouped in anROI and those, which fall below, are discarded ormade transparent. The frequency distributiongraph for the fossil shark tooth reveals three peaksrepresenting air, mineral-filled spaces and rockmatrix as well as the fossil (Figure 3). The fossilROI was defined by applying a global threshold atthe grey value minima that separated the matrixpeak from the fossil (Figure 3). The calibration wasconverted into an ROI (Figure 4.1 - cyan bound-ary). ROIs can be renamed in a meaningful way; inthis case one ROI was created for the voxels rep-resenting any biologically mineralized tissues thatmake up the tooth (dentine and enameloid). How-ever, since there were also fragments from otherteeth embedded in the matrix, the tooth region ofinterest automatically included more than one fossil(Figure 4.1 - cyan boundary). This ambiguity canbe resolved at a later stage (see below).

Region Growing

The fossil tooth was separated from otherfragmentary remains by incorporating informationregarding location. This separation was achievedusing a 3D region-growing (or magic wand) tool.Region growing is an image processing techniquethat segments all voxels that are connected to aseed point and that are within a user defined toler-ance (visually of same density) of voxel grey val-ues (see Ronan et al., 2010). The larger thetolerance the more voxels will be selected. Bysearching only within the calibrated ROI it was pos-sible to remove all non-adjacent teeth and frag-ments. Wand tools are most successful when theseed represents a low intensity voxel from the fos-sil (i.e., enameloid). By searching within the cali-brated ROI (Figure 4.1 - cyan boundary) using aseed value of 45571 and a threshold of +20000 itwas possible to create a new region of interest(Figure 4.2 - yellow boundary), which excludedsome extraneous dental fragments (Figure 4.2 -cyan boundary). However, some adjacentunwanted material remained linked to the ROI.

Masking

Together, surface determination and regiongrowing are not usually quite sufficient to com-pletely prepare a fossil. Some regions of matrix oradjacent fragments must be manually removedusing a process referred to as masking. This is typ-ically the most time consuming, subjective and

FIGURE 3. Micro-CT grey value frequency distributionplot. The graph reveals three peaks representing air,matrix and fossil. The tooth was calibrated by applying aglobal threshold at the minima, which separated the fos-sil peak (vertical red line).

FIGURE 4. Virtual preparation (segmentation) of thefossil was carried out in three steps. (4.1) Surface deter-mination (4.2) ROI growing and (4.3) masking. (4.4)The vascular system was extracted using a combinationof masking and region growing.

6


(usually) unavoidable aspect of segmentation. Vox-els are manually added or removed from an ROIslice by slice (up to 2000 of them) or sets of fewslices using a drawing or eraser tool (Figure 4.3 -green area). In our example the calibrated andregion- grown ROI (Figure 4.2 - cyan boundary)was masked using an eraser tool (Figure 4.3 -green area) to produce a final, virtually preparedfossil, which could be dissected (Figure 4.4).

Virtual Dissection

An ROI representing the vascular system ofthe tooth was dissected using a combination of theregion growing and masking tools. Voids repre-senting the canals were extracted using the wand.The voids were continuous with the rock matrixthough because of mineral infillings during the fos-silization. Hence it was necessary to block (mask)the openings of the dental foramina. The foraminawere blocked in 3D by manually creating smallregions of interest slice by slice using a drawingtool (Figure 4.4 - green ROI). The wand tool wasinstructed to search within vascular canals in 3Dfrom a starting seed point inside the void with a rel-atively pixel value of 32666, and the wand toler-ance was set at +20000. The wand was instructedto ‘avoid’ other defined regions of interest, i.e., fos-sil tooth and blocking mask. Hence the tool onlyselected the vascular canals without spilling out-side the fossil tooth (Figure 4.4 - red ROI). The cre-ation of a region of interest produced a 3Dsegmentation of the vascular network, which couldbe visualised independently from the rest of thetooth (Figure 5).

In this case study, the virtual reconstructionreveals in great detail the external features of thetooth, as well as the internal dental anatomy (seeFischer et al., 2010), including most of the mor-phology of the vascular system (Radinsky, 1961),the type of dentine (trabecular versus ortho den-tine) and the distribution of the enameloid cover onthe crown (see Botella et al., 2009). Hence the vir-tual specimen reveals more information about thetooth than a mechanically prepared specimenwould, even if it was sampled destructively by sec-tioning or chemical preparation, as the reconstruc-tion and gross histological features are shown in3D. However, the virtual dissection includes cracksin the tooth that are continuous with the vascularcanals. Consequently there are some errors in thefinal prepared and dissected tooth (Figure 5).Given more time, these errors could be corrected.Users need to decide whether further segmenta-tion is required given the scientific questions at

hand or the desired quality of the virtual prepara-tion.

HOW CAN I VISUALISE AND SHARE A VIRTUAL FOSSIL?

A CT scan is essentially a matrix made up ofgrey-scale voxels. The segmentation process sim-ply groups the voxels together and defines them asmaterials/phases/tissues etc. The first segmenta-tion and virtual preparation of fossils was relativelyrecent (Kearney et al., 2005; Balanoff and Rowe,2007) but the technique is rapidly becoming a stan-dard in palaeontology. In order to visualise a whole

FIGURE 5. Virtually prepared fossil rendered with lights,colour and perspective. (5.1) Fossil and vascular sys-tem, (5.2) transparent fossil and vascular system and(5.3) vascular system.

7


fossil or the segmented anatomical features thevoxel data must be imaged or animated. In thesame way that one might create photographic stillsor videos of the original fossil. It is worth noting thatthe protocols discussed here can also be applied toprocess any voxel based datasets produced using3D imaging modalities such as MRI, synchrotrontomography and confocal microscopy.

Volume Rendering

Rendering is the process of generating animage from a 3D segmented volume using com-puter software. The segmented tooth and vascularsystem were rendered using lights, colour, stereo-perspective (Figure 5) and stereo-anaglyphy (Fig-ure 6). Firstly the colour of the fossil was selectedto resemble the original specimen, and the vascu-lar system was reddened. Secondly, 3D perspec-tive was applied to the virtual fossil by making themore distant parts of the object smaller than thoseclose to the viewer, i.e., spatial shortening. Thirdly,creating shadows enhanced the sense of perspec-tive. Two directional light sources were simulated;an ambient light directed straight at the specimenand a spotlight angled from the upper left, as iscustomary in photography (Figure 5.1). The exactposition of the lights was adapted to match eachrendering of the tooth/vascular system and set thesurface markings or vascular system in clear con-trast (Figure 5.2-3). The manner in which lightreflected off the ‘virtual’ fossil tooth was modelledusing the Phong algorithm (Phong, 1975) whererough and smooth surfaces reflected light in specu-

lar and diffuse patterns respectively. Finally, a ste-reo-anaglyph was created to provide a fullstereoscopic 3D effect, when viewed with red-cyanspectacles (Figure 6, Supplementary File 2).

Movie Animation

Animation is an excellent tool for sharing data-sets and can be used to draw the eye of the endviewer to specific features of interest. Movies alsoallow a degree of interactivity with the user able tovisualise a fossil from different angles and perhapsmake independent interpretations about the mor-phology (see Parfitt et al., 2010, Abel et al., 2011).In order to demonstrate the quality and usefulnessof the virtual fossil, an animation, rotating onceabout the long axis, was created (see Supplemen-tary File 1). The simple animation depicts a cameraflying around the tooth whilst the fossil becomestransparent to reveal the vascular canal systeminside (see Supplementary Files 1, 2). This effectwas achieved using key framing: a techniquebased on traditional hand-drawn animation. Senioranimators would draw the key frames from an ani-mation sequence then a junior assistant woulddraw the frames in between. Computerised key

FIGURE 6. Stereo-anaglyph of virtually prepared fossil.Rendered with lights, colour and perspective. The imagecan be viewed in three-dimensions using red-cyan (orred green) spectacles. FIGURE 7. Animation keyframer and key frames used

to produce supplementary movies 1 and 2. The camera(yellow) follows a circular path (green) around the tooth(shown as a cross section). At each key frame the cam-era is rotated toward the specimen. At key frames 1 and2 the tooth is opaque but at 3, 4 and 5 it is transparent.The key frame interpolates these settings as a singlecamera rotation about the tooth with a transparencydeveloping between key frames 2 to 3. This reveals theunderlying vascular structure.

8


framer applications work in a similar way by inter-polating an animation across user-defined keyframes or camera views. In this case, five evenlyspaced key frames were defined around the speci-men, each frame was 45º apart, except the firstand last frames, which were both located at 0º(Figure 7). Thus the movie starts and ends at thesame position. At each key frame the renderingsettings were defined, e.g., camera angle, lightingand tooth transparency. The key framer then auto-matically interpolated the complete path of thecamera and rendering settings to create the 360ºfly around. The videos are small enough to email(<16Mb) and can be embedded in PDF files oradded to publications as downloadable supple-mentary information.

Stereolithographic Files

It is possible to create even smaller files byconverting CT datasets to stereo-lithographic (STL)files. An STL file is a surface mesh (usually trian-gles), which describes both the internal and exter-nal geometry of a three-dimensional object.Creation of an STL file is akin to physically stretch-ing a triangular mesh over the voxels in the virtualmodel creating a smooth appearance of the voxelsand slicing them into a continuous surface. Exam-ple STL files of the tooth (Supplementary File 3)and vascular system (Supplementary File 4) werebuilt. The files can be opened up in many 3D imag-ing software packages (see Table 2). Users canmanipulate, rotate and section STL files. Hencethey are less restricted and more interactive thananimated movies, but end users require bettercomputer software skills.

WHICH SOFTWARE SHOULD I USE TO CREATE ‘VIRTUAL’ FOSSILS?

A huge number of commercial and freewarepackages are available for segmentation and ren-dering CT data. The most widely used packagesare listed in Table 2 along with a synopsis of func-tionality and a list of URLs. The ability of the pro-grammes to accomplish the key CT dataprocessing steps varies. The data processingsteps incorporated within each programme arehighlighted in Table 2. Ultimately each CT usermust select appropriate software based on the par-ticular data processing tasks required.

We have found that generally the most usefulsoftware packages for processing scans of fossilare VGStudio MAX 2.1 (Volume Graphics, Heidel-berg, Germany) SPIERS (Selden et al., 2008; Gar-wood et al., 2009) and Drishti (Jones et al., 2007;

Sakellarioua, 2007). For this paper the shark toothCT data was post segmented and rendered usingVGStudio MAX 2.1 because the programmeincludes all of the data processing tools discussedhere. Since the programme can carry out thresh-olding and segmentations in 3D using surfacedetermination and wand tools the speed at whichfossils can be segmented is faster in comparison tomost other software packages. However, the basepackage costs around £9K, but is available for lesswith academic discounts. SPIERS is a simple buthighly capable programme for slice based (i.e., 2D)masking segmentation of CT data making the pro-gramme ideally suited to many fossils, particularlythose with low density contrast scans. In contrast,Drishti is not as useful for thresholding or segment-ing CT data and is rather restricted to basic imagecontrast enhancement. However, the programmeis excellent for rendering movie animations, spinsor flythrough, of fossil specimens. Alternativelyexcellent movies can be produced using opensource software such as SPIERS to create STLfiles then animating them using, e.g., BLENDER.Ultimately advising users which software packagesto adopt is not straightforward. Users’ requirementsare often esoteric but SPIERS is a useful (free)place to start and is well supported by the authors(Garwood et al., 2009; Garwood et al., 2010; Gar-wood and Sutton, 2012; Sutton et al., 2012).

HOW DO I MEASURE A ‘VIRTUAL’ SPECIMEN?

Sphenacanthus was originally CT-scannedbecause of the scientific interest in describing itsinternal dental anatomy and gross histology, suchas the type of dentine that makes up the crown andbase of the tooth. The voxel based micro-CT scanis a mathematical representation of an object,rather than an image, which contains accuratescale information and is therefore suited to thequantitative analysis of geometry and structure in3D. For example, dental crown morphology hasbeen characterised using landmark-based geomet-ric morphometric measures collected directly fromCT scans (Gomez-Robles, 2007, 2008). There aremany programmes for collecting landmark datasuch as Stratovan Checkpoint (see Table 2). Simi-larly the structure of the vascular network could bemeasured using ImageJ (Abramoff, 2004) pluginssuch as BoneJ (Doube et al., 2010). Originallyintended for analysing bone, the software can cal-culate the thickness, connectivity and orientation oftrabeculae (cancellous bone). However, the soft-ware could be co-opted for measuring vascular

9


structure. For a description of how the measuresare calculated visit BoneJ.Org.

However, it is important to note that the accu-racy of measurements collected from CT Scanswill be affected by the quality of the scanning andsegmentation. Scanning artefacts such as partialvolume averaging, noise and beam hardeningobscure material boundaries within a CT scan.Hence, segmentations based on surface determi-nation, region growing or masking will not beentirely accurate. Furthermore, segmentation is asubjective process. Selection of global thresholds,wand seed points and tolerance varies acrossusers. Similarly manual masking is a subjectiveprocess, which relies heavily on the prior anatomi-cal knowledge of the user. Therefore, users shouldassume that segmentations are subject to errorand (where possible) avoid measuring featuresclose to or smaller than the resolution of the scan.As a rule of thumb this would be no smaller thantwo to five times the voxel size. Scan resolutioncan be determined accurately by scanning andmeasuring plastic (TEM) grids with meshes ofvarying size. The grids can be scanned alongside afossil or separately, the finest grid that can be visu-

alised and measured is equal to the resolution ofthe scan.

HOW CAN I DISSEMINATEVIRTUAL COLLECTIONS?

Quantitative analysis of fossil geometry canrequire large comparative samples. Museums anduniversities are starting to build up and share sig-nificant ‘virtual collections’ of specimens. The firstattempt to systematically distribute fossil CT datawas made by Rowe and colleagues (Rowe et al.,1993, 1995). The reptile fossil Thrinaxodon wasscanned and slice stacks made available on CD.This was a landmark in CT and more generally inthe Natural Sciences. Since then the Internet hasallowed researchers to share data more directly, forexample via the inspeCT applet on DigiMorph.Other organisations such as NESPOS and the Nat-ural History Museums (London, UK) are also shar-ing CT data. However, the advancement of CTbased morphological research is being held backby a lack of significant data archives available fordownload and visualisation at full resolution (seeRowe and Frank, 2012), even though the technol-

TABLE 2. Data processing software for voxel and surface based 2D and 3D data. The useful tools and applications

included in the packages are highlighted.

Package Source Su

rfac

e D

eter

min

atio

n

Mag

ic W

and

Mas

kin

g

Ren

der

ing

Ste

reo

-An

agly

ph

s

Mo

vie

An

imat

ion

ST

L F

iles

Tra

dit

ion

al M

orp

ho

met

rics

Ge

om

etr

ic M

orp

ho

met

rics

URL

VG Studio Max Commercial 3D 3D 2D 3D 3D 3D 3D 3D 3D www.volumegraphics.com

SPIERS Freeware 3D 2D 3D 3D www.spiers-software.org/

DRISHTI Freeware 3D 3D 3D www.sf.anu.edu.au/Vizlab/drishti/

ImageJ Open Source 2D 2D 3D 3D 2D www.rsb.info.nih.gov/ij/

Mimics Commercial 3D 2D 3D 3D 3D www.materialise.com/mimics

Simpleware Commercial 2D 2D 2D 3D 3D 3D 3D www.simpleware.com

Amira Commercial 3D 2D 2D 3D 3D 3D www.amira.com/

Blender Open Source 3D 3D www.blender.org

Stratovan Checkpoint

Commercial 3D 3D www.stratovan.com

Meshlab Commercial 3D 3D www.meshlab.sourceforge.net/

Simpleware Commercial 2D 2D 2D 3D 3D 3D 3D www.simpleware.com

10


ogy and software for sharing data sets across theInternet is already available to most users.

Segmented or un-segmented datasets can beeasily disseminated across the Internet in a fewhours with the use of ftp sites. Provided that aSecure File Transfer Protocol (SFTP) encryption isused, the data can be transported safely (Abel etal., 2011). The file size of CT datasets may be thefactor limiting storing and dissemination of CTdatasets. A set of projections from a CT scan canbe 25 GB in size, and the reconstructed 3D volumecan be 32 GB. Consequently data sets are oftendownsized from 32/16-bit to 8-bit, and although thecompression reduces contrast density this is usu-ally sufficient for most users’ needs. Particularlywhen there are only three phases, e.g., air, matrixand fossil. Alternatively, data can be stored asJPEG2000 format, which has a very high compres-sion with minor loss of information. However, creat-ing and uncompressing 32 GB of JPEG 2000 filescan take 30 minutes or more.

STL files can be emailed or file exchangedvery quickly or added as supplementary informa-tion to a paper. Typically STLs are only a few tensor hundreds of megabytes in size. An STL file canalso be rapidly prototyped to produce a 3D plasticmodel (see Abel et al., 2011). Models are builtusing 3D printers by tracing out a stack of CTcross-sections in layers of UV-curable resin. Thismirrors the creation of a CT volume from a slicestack to a great extent. The resolution of the mod-els is usually between 0.5 to 0.1 mm, which is usu-ally high enough to capture fine surface topology.The resin models can be scaled up or down for usein education and research, e.g., in museum galler-ies, schools and universities. When conceptualis-ing the size and shape of fossils, many users,particularly students and school children, may findthat 3D prints are more useful than virtual models.

DISCLAIMER: IS MICRO-CT A ‘SILVER BULLET’?

The majority of fossils cannot be scannedusing current CT technology. Palaeobiologistsmust carefully consider whether a given fossil issuitable for CT, or whether a more traditional tech-nique would be more useful. There are various rea-sons why fossils cannot be scanned. Firstly, muchof the paleontological record (estimated at ~90%)consists of specimens flattened by the fossilisationprocess. Flat objects are usually very difficult to CTscan because of the unequal path length for x-rays.The broadside is often overexposed while theedges are underexposed.

Secondly, for 3D fossils, size and density arethe main limiting factors. A scanner must be largeenough to accommodate a specimen and powerfulenough to produce X-rays that can penetrate thematerial. Most commercially available scanners donot have a field of view greater than 250 mm indiameter and height (e.g., HMX-ST CT System,Nikon Metrology, Tring, UK). Furthermore, mostsystems can only generate X-rays with energy ofbetween 20-450 kV. At maximum power a 225 kVsystem can penetrate a block of limestone 100-200mm in diameter whilst a 450 kV system can pene-trate a lump of ironstone, which is 50-100 mmacross.

Thirdly, in order for the scanner to distinguishthe materials in a scanned fossil, e.g., bone andmatrix, there must be sufficient density contrastbetween them. Where two materials have similardensities, the voxels that represent the tissue willhave comparable grey values to the host rock. Ifthis is the case, blurring caused by partial volumeaveraging effects is compounded, and noise pre-vents segmenting. For equivalent voxel size at low-density contrast, such as a calcite fossil in lime-stone, the blurring is greater than at high-densitycontrast such as a fossil void in siderite. Enhance-ment can improve grey value contrast between twomaterials, but tends only to be useful when there isalready reasonable contrast, and the user wants tocreate a sharper cutoff for masking (manual) seg-mentation (Dominguez et al., 2002).

It is also worth mentioning that CT is not asubstitute for other techniques aimed at resolvingmorphological features of biological materials atfine histological level. An example is light micros-copy of thin-sectioned specimens. Although actual(as opposed to virtual) thin-sectioning is a ‘destruc-tive’ sampling technique, it can be justified andnecessary in a number of cases, for instance whenCT cannot resolve structures such as cell spaces,dentinal tubules or Sharpey fibers. Although someCT systems are capable of resolving features atthe nano-scale (e.g., nano-CT, Synchrotron andXuM).

CONCLUSONS: WHAT NEXT FOR CT?

Micro-CT is a powerful non-destructive imag-ing modality for the full-volume visualization of boththe external and internal aspects of an object in3D. Hence it has become very popular in biologicalresearch, allowing the rich detail of internal struc-tures to be retrieved, without any damage to thespecimens (see Elliot and Dover, 1982; Flannery etal., 1987; Rowe et al., 1997). However, as seen

11


above, computed tomography is not a suitabletechnique for all fossils. Scanning is limited by theoverall size and density of a fossil as well as thedensity contrast between phases, e.g., matrix andbone, and relies on three-dimensional preserva-tion. Where applicable virtual preparation doeshave several advantages over mechanical prepa-ration. Digital preparation is quicker than manymechanical or chemical preparations. The digitalprocess requires less training and experience,approximately 1-2 days of training. Few researchinstitutions and museums can afford to employ fos-sil preparers, who develop their highly specializedprofessional skills through many years of experi-ence.

Furthermore, the original fossil and hostmatrix context is not destroyed or damaged in anyway through CT scanning. Since fossils are notalways completely visible at the start of the prepa-ration process, it is possible to make mistakes thatcan damage the specimen. Every mechanical oracid preparation of fossils is therefore a uniqueexercise that is also usually very time consumingand costly. The original CT data can be curatedalong with actual specimens, providing a perma-nent digital record of the specimen. Should prac-tices change or imaging processing softwareimprove, then virtual preparations can be repeatedutilising the original scan. Presently the CT tech-nique is still waiting for further innovations that willbring the technology into mainstream research.This technology would be attributable to three fac-tors, affordability, productivity and computingpower.

Firstly, systems and scans are becomingmore affordable for museums and universities soat some point every institution will have CT andcomputer laboratories as standard. Secondly, scantime and field of view, which limit the number andsize range of specimens that can be scanned, areconstantly improving. Some systems such as theHMXST CT (Nikon Metrology, Tring, UK) canalready collect a scan in 10-20 minutes. New sys-tems are being developed incorporating automaticscanning and specimen changes. Allowing sys-tems to run over weekends and holidays will alsogreatly increase CT capacity. The field of view isalso increasing. The size of detector panels dou-bles every two years, and new software algorithmsallow scans to be stitched together. Using theHMX-ST system, which has a field of view 250x250mm, it is possible to scan an object 1000x250 mmin four sections and stitch them together virtually.

At the present some manual labour and patienceon the part of the user is required but sooner orlater this function will be automated. Hence, muse-ums and universities will be able to build up signifi-cant ‘virtual collections’ from a varied array ofspecimens and materials. Thirdly, computingpower is increasing dramatically relative to cost.Until recently, computing power lagged way behindCT scanning technology. It is only in the last fiveyears that computer prices have fallen enough foran increasing number of palaeobiologists to invest.At present a PC suitable for segmenting and ren-dering will require about 2TB of storage and 64 GBof RAM but costs only £5K. However, in order forcollaborators to share large datasets on powerfulcomputers we will need faster Internet speeds.

Despite these ongoing and future develop-ments, we still have a long way to go in order tomake entire museum or university collections avail-able to scientists and the public alike in digital form.The Natural History Museum in London alone con-tains approximately 75 million specimens, whichamounts to generations worth of scanning, recon-struction, preparation, dissection, rendering, mea-surement and quantitative analysis. However, if wecould routinely exploit CT and other 3D imagingtechniques to share data more widely either freelyor at a low cost, we could be ushering in a new eraof global scientific collaboration and public commu-nication. CT scanning is an unbeatable techniquefor discovering hidden fossils inside rocks andinternal features of specimens that until now havebeen impossible to visualise without destroyingthem.

ACKNOWLEDGMENTS

The authors would like to thank: R. Garwood(The University of Manchester); T. Rowe (The Uni-versity of Texas at Austin); A. Ramsey (NikonMetrology, Tring, UK) and D. Handl (VolumeGraphics, Heidelberg Germany) for helpful com-ments and advice as well as A. Ball and L. Howard(Natural History Museum, London).

REFERENCES

Abel, R.L., Parfitt, S.A., Ashton, N.M., Lewis, S.G., andStringer, C.B. 2011. Digital preservation and dissemi-nation of ancient lithic technology with modern micro-CT. Computers and Graphics, 35:878–884.

Abramoff, M.D., Magelhaes, P.J., and Ram, S.J. 2004.Image processing with ImageJ. Biophotonics Interna-tional, 11:36-42.

12


Balanoff, A.M. and Rowe, T.B. 2007. Osteologicaldescription of an embryonic skeleton of the extinctElephant bird, Aepyornis (Palaeognathae, Ratitae).Memoir 9, Society of Vertebrate Paleontology, Jour-nal of Vertebrate Paleontology, Supplement to 27:1-54, plus supplementary CD-ROM.

Botella, H., Donoghue, P.C.J., and Martinez-Perez, C.2009. Enameloid microstructure in the oldest knownchondrichthyan teeth. Acta Zoologica, 90:103-108.

Cline, H.E., Lorensen, W.E., Ludke, S., Crawford, C.R.,and Teeter, B.C. 1998. Two algorithms for the threedimensional construction of tomograms. MedicalPhysics, 15:320-327.

Croft, W.N. 1950. A parallel grinding instrument for theinvestigation of fossils by serial sections. Journal ofPaleontology, 24:693-698.

Dick, J.R.F. 1998. Sphenacanthus, a Palaeozoic fresh-water shark. Zoological Journal of the Linnean Soci-ety, 122:9-25.

Dominguez, P., Jacobson, A.G., and Jefferies, R.P.S.2002. Paired gill slits in a fossil with a calcite skele-ton. Nature 417, 841-844.

Doube, M., Kłosowski, M.M., Arganda-Carreras, I., Cord-elières, F., Dougherty, R.P., Jackson, J., Schmid, B.,Hutchinson, J.R., and Shefelbine, S.J. 2010. BoneJ:free and extensible bone image analysis in ImageJ.Bone 47:1076-9.

Egerton, P.M.G. 1853. Palichthyologic Notes No 5 - Ontwo new species of placoid fishes from the Coal Mea-sures. Quarterly Journal of the Geological Society ofLondon, 9:280-282.

Elliott, J.C. and Dover, S.D. 1982. X-ray microtomogra-phy. Journal of Microscopy, 126:211-213.

Fischer, J., Schneider, J.W., and Ronchi, A. 2010. Newhybodontoid shark from the Permocarboniferous(Gzhelian-Asselian) of Guardia Pisano (Sardinia,Italy). Acta Palaeontologica Polonica, 55:241-264.

Flannery, B.P., Deckman, H.W., Roberge, W.G., andD’Amico, K.L. 1987. Three-dimensional X-ray micro-tomography. Science, 237:1439-1444.

Garwood, Russell J. and Sutton, Mark D., 2012. Theenigmatic arthropod Camptophyllia. PalaeontologiaElectronica Vol. 15, Issue 2;15A,12p; palaeo-electronica.org/content/2012-issue-2-articles/218-the-arthropod-camptophyllia

Garwood, R.J., Dunlop, J.A., and Sutton, M.D. 2009.High-fidelity X-ray micro tomography reconstructionof siderite-hosted Carboniferous arachnids. BiologyLetters, 5:841-844.

Garwood, R.J., Rahman, I.A., and Sutton, M.D. 2010.From clergymen to computers – the advent of virtualpalaeontology. Geology Today, 26:96-100.

Gomez-Robles, A., Martinon-Torres, M., Bermudez deCastro, J.M., Prado, S., Sarmiento, S., and Arsuaga,J.L. 2008. Geometric morphometric analysis of thecrown morphology of the lower first premolar of hom-inins, with special attention to Pleistocene Homo.Journal of Human Evolution, 55:627-638.

Gomez-Robles, A., Martinon-Torres, M., Bermudez deMargvelashvili, A., Bastir, M., Arsuaga, J.L., Perez-Perez, A., and Martinez, L.M. 2007. A geometricmorphometric analysis of hominin upper first molarshape. Journal of Human Evolution, 53:272-285.

Hounsfield, G.N. 1973. Computerized transverse axialscanning (tomography): part I. Description of system.British Journal of Radiology, 46:1016-22.

Jones, A.C., Arns, C.H., Sheppard, A.P., Hutmacher,D.W., Milthorpe, B.K., and Knackstedt, M.I. 2007.Assessment of bone in growth into porous biomateri-als using Micro-CT. Biomaterials, 28:2491-2504.

Kearney, M., Maisano, J.A., and Rowe, T. 2005. Cranialanatomy of the extinct worm-lizard Rhineura hatch-erii (Squamata, Amphisbaenia) based on High-Reso-lution X-ray Computed Tomography. Journal ofMorphology, 264:1-33, plus Web supplement atwww.DigiMorph.org.

Maisey, J. 1982. Studies on the Paleozoic selachiangenus Ctenacanthus Agassiz: No. 2. BythiacanthusSt. John and Worthen, Amelacanthus, new genus,Euemacanthus St. John and Worthen, Sphenacan-thus Agassiz, and Wodnika Münster. AmericanMuseum Novitates, 2722:1-24.

McColl, D.J., Abel, R.L., Spears, I.R., Macho, G.A. 2006.Automated method to measure trabecular thicknessfrom microcomputed tomographic scans and itsapplication. Anatomical Record. Part A, Discoveriesin Molecular, Cellular and Evolutionary Biology,288:982-988.

McCormack, A.M. 1963. Representation of a function byits line integrals, with some radiological applications.Journal of Applied Physics, 34:2722-2727.

Parfitt, S.A., Ashton, N.M., Lewis, S.G., Abel R.L.,Coope, G.R., Field, M.H., Gale, R., Hoare, P.G., Lar-kin, N.R., Lewis, M.D., Karloukovski, V., Maher, B.A.,Peglar, S.M., Preece, R.C., Whittaker, J.E., andStringer, C.B. 2010. Early Pleistocene human occu-pation at the edge of the boreal zone in northwestEurope. Nature, 466:229-233.

Phong, B.T. 1975. Illumination for computer generatedpictures. Communication of the ACM, 6:311-317.

Radinsky, L. 1961. Tooth histology as a taxonomic crite-rium for cartilaginous fishes. Journal of Morphology,109:73-92.

Reissis, D., and Abel, R.L. 2012. Development of fetaltrabecular micro-architecture in the humerus andfemur. Journal of Anatomy, 220:496-503.

Ronan, R., Abel, R.L., Johnson, K., and Perry, C. 2010.Quantification of porosity in Acropora pulchra (Brook1891) using X-ray micro-computed tomography tech-niques. Journal of Experimental Marine Biology andEcology, 396:1-9.

Ronan, R,. Abel, R.L., Johnson, K., and Perry, C. 2011.Spatial variation in porosity and skeletal elementcharacteristics in apical tips of the branching coralAcropora pulchra (Brook, 1891). Coral Reefs,30:195-201.

13


Rowe, T. and Frank, L.R. 2012. The disappearing thirddimension. Science, 331:712-714.

Rowe, T., Carlson, W., and Bottdorf, W. 1993. Thrinax-odon Digital Atlas of the Skull, CD-ROM. Universityof Texas Press, Austin, 623 megabytes.

Rowe, T., Carlson, W., and Bottorff, W. 1995. Thrinax-odon: Digital Atlas of the Skull. CD-ROM (SecondEdition). University of Texas Press, 547 megabytes.

Rowe, T., Kappelman, J., Carlson, W.D., Ketcham ,R.A.,and Denison, C. 1997. High-resolution computedtomography: a breakthrough technology for earth sci-entists. Geotimes, 42:23-27.

Sakellarioua, A., Arnsa, C.H., Shepparda, A.P., Soka,R.M., Averdunka, H., Limaye, A., Jones, A.C., Sen-den, T.J., and Knackstedt, M.A. 2007. Developing avirtual materials laboratory. Materials Today, 10:44-51.

Selden, P.A., Shear, W.A., and Sutton, M.D. 2008. Fossilevidence for the origin of spider spinnerets, and aproposed arachnid order. Proceedings of NationalAcademy of Sciences USA, 105:20781-20785.

Sollas, W.J. 1904. A method for the investigation of fos-sils by serial sections. Philosophical Transactions ofthe Royal Society of London. Series B, BiologicalSciences, 196: 259-265.

Spoor, C.F., Zonneveld, F.W., and Macho, G.A. 1993.Linear measurements of cortical bone and dentalenamel by computed tomography: applications andproblems. American Journal of Physical Anthropol-ogy, 91:469-484.

Sutton, Mark D., Garwood, Russell J., Siveter, David J.,and Siveter, Derek J. 2012. SPIERS and VAXML; Asoftware toolkit for tomographic visualisation and aformat for virtual specimen interchange. Palaeontolo-gia Electronica Vol. 15, Issue 2;5T,14p; palaeo-electronica.org/content/94-issue-2-2012-technical-articles/226-ct-toolkits

Xu, F. and Mueller, K. 2005. Accelerating popular tomo-graphic reconstruction algorithms on commodity PCgraphics hardware. IEEE Transactions on NuclearScience, 52:654-663.

Zonneveld, F.W. 1987. Computed tomography of thetemporal bone and orbit. Munich, Urban andSchwarzenberg.

14


GLOSSARY OF CT TERMS

Beam Hardening Phenomenon referring to theeffect of selective x-ray attenu-ation and scatter from poly-chromatic X-ray beams. An x-ray beam is composed of indi-vidual photons with a range ofenergies. As the beam passesthrough an object, it becomes"harder" (mean energyincreases) because the lower-energy photons are selec-tively absorbed (or scattered)more rapidly, leaving behindonly the high-energyphotons. Both cupping andstreak artefacts can occur as aresult.

Cupping artifact A ring of excessively brightvoxels around the edge of thefossil reducing to excessivelydark voxels at the centre –caused by beam hardening

CT numbers Term coined by GodfreyHounsfield to describe thevoxel grey values

Electron volt (eV) is a unit of energy equal to theamount of kinetic energygained by a single unboundelectron when it acceleratesthrough an electric potentialdifference of one volt.

LUT plot Image look up table, a bivari-ate plot of grey scale voxel val-ues (CT numbers) versus thenumber of voxels

Noise Unusually bright and dark vox-els are superimposed over theCT data, which gives a CTscan a grainy or speckledappearance, leading to blur-ring of material boundaries.

Partial volume Artefact resulting when numer-ous linear attenuation coeffi-cients (i.e., material density)different within a single voxelbeing represented by an aver-aged grey value.

Segmentation The process of definingregions of interest with a scan,which usually represent a fos-sil or part thereof. The regionsof interest can be modelledand rendered independently.Segmentation invariablyrequires the user to threshold(i.e., define) material boundar-ies within an object e.g., fossiland matrix (note: the term seg-mentation refers to a differentprocess in the SPIERS soft-ware suite).

Polychromatic An X-ray beam composed ofradiation of more than onephoton energy (i.e., wave-length)

Steak artefacts Dark bands or streaksbetween dense objects in taCT slice image.

Threshold The process of defining mate-rial boundaries within a scan.In its simplest form this entailsselecting a grey value (thresh-old) above which any voxelsare grouped in a region ofinterest and those, which fallbelow, are discarded or madetransparent.

Reconstruction is the process of transformingthe set of 2D X-ray projectionsto a 3D volume.

Region of interest (ROI) is a selected sub-regionof a CT scan that can be iden-tified e.g., a fossil tooth withina matrix or vascular voids withthe tooth.

Rendering the process of generating animage from a model

Surface determination defining the surface of anobject to produce a hollow 3Dmodel without internal features

15


APPENDIXES

SUPPLEMENTARY FILE 1. Virtual fossil animation, rotating once about the long axis and revealing thevascular structure. The specimen is rendered with lights, colour and perspective.

SUPPLEMENTARY FILE 2. Stereo-anaglyph virtual fossil animation, rotating once about the long axis andrevealing the vascular structure. The specimen is rendered with lights, colour and perspective. The moviecan be viewed in three-dimensions using red-cyan (or red green) spectacles.

SUPPLEMENTARY FILE 3. Stereolithographic (STL) model of the fossil shark tooth. The 3D model can beopened by readers using freeware (see text). and SUPPLEMENTARY FILE 4. Stereolithographic (STL)model of the dental vascular system. The 3D model can be opened by readers using freeware (see text)(zipped file online).

16

Documents

Palaeontologia Electronica · Palaeontologia Electronica palaeo-electronica.org ... (Egerton), a shark from the Late ... rare specimens and type materials in university and museum