Tomography software

From NoskeWiki
Jump to navigation Jump to search


A "tomogram" is a series of 2D images, called "slices" which stack together to form a 3D image. There are many types of tomography, and I work with one called electron tomography. There are many programs designed to view tomograms, and I have listed some of the more popular choices below - particularly programs which can segment and visualize the data.

Free Segmentation Programs


  • Website/Ref: IMOD (Kremer et al. 1996)
  • Cost: FREE
  • Platform: ALL PLATFORMS
  • About: Used to reconstruct tomograms from tilt series, and then segment tomograms by drawing contours to produce an IMOD model file. Consists of the 3dmod (Qt/C++), eTomo (Java) and Midas (Java) GUI, but also has numerous command line programs, including a couple of different analysis tools. Has some good filtering options, but can only make movies linearly from A to B. Via the main program author (David) I was able to add my own plugins, DrawingTools, Interpolator and BeadHelper onto this program (Noske 2010).
  • Developed: University of Colorado (Boulder). Main author is David Mastronarde


  • Website/Ref: UCSF Chimera homepage (Pettersen et al. 2004)
  • Cost: FREE
  • Platform: ALL PLATFORMS. Written in Python and C++
  • About: Chimera is primarily used for looking at molecular structures and single-particle EM maps, but its ET capabilities are always improving and it does a terrific job of fast image processing and generating isosurfaces over density maps etc. Can load IMOD image and model files, image stacks and numerous other formats. Well worth looking at the image gallery and animation gallery to see what is possible!
  • Developed: University of San Francisco. Main Author is Tom Goddard and PI Tom Ferrin

ImageJ + TomoJ + NeuroJ + Fiji + TrakEM2

  • Website/Ref: ImageJ (Messaoudi et al. 2007), TomoJ (Collins 2007), NeuroJ (Meijering et al. 2004 [1]), Fiji, TrakEM2 (Cardona et al. 2010, [2])
  • Cost: FREE
  • Platform: ALL PLATFORMS
  • About: ImageJ is an open source java-based image processing program designed for analysis of various microscope data - you can read more about it on my wiki page ImageJ. ImageJ is open source and encourages plugins. TomoJ is a tomography plugin for TomoJ. NeuroJ is a plugin for tracing neurons, but only works on 2D images.Fiji is a easy-to install version of ImageJ which is simply ImageJ with Java 3D and a large number of plugins already pre-installed. For example good plugin for local autocontrasting is: Clathe. More recently, they've added TrakEM2 which is a large suite of proper segmentation and visualization tools specific for EM tomograms and even include some nice video tutorials.
  • Developed: Various - ImageJ is developed by an NIH employee, Wayne Rasband (see Wikipedia - ImageJ) who is based in Washington DC (I believe), while Fiji and TrackEM are based in Europe.


  • Website/Ref:
  • Cost: FREE with permissive license.
  • Platform: ALL PLATFORMS. Written in C++
  • About: Population is dedicated to the quantitative/qualitative analysis of images coming from 2D/3D microscopy with filtering, segmentation, geometrical/physical characterization, visualization and modeling. Its integration in the caméléon language allows to prototype a work-flow graphically as a Lego game.
  • Developed: Ecole Polytechnique. Main Author is Vincent Tariel and Olivier Cugnon de Sevricrout


  • Website/Ref: Reconstruct (Filala et al. 2005 [3])
  • Cost: FREE
  • Platform: Windows only.
  • About: Reconstruct is very similar to 3dmod/IMOD: it's able to reconstruct slices, montage, segment etc and even uses the same shortcut keys/methods as IMOD. One feature which makes it interesting/different is that that "original image data are never altered" (every section and every "trace" has a transform associated and displayed in real time), however this can also make it slow. I found segmentation using this tool painful, and the "wildfire growing tool" was a nice idea but very ineffective! This program is closely associated with two other Boston programs called "IGL Trace" and "sEM Align".
  • Developed: Boston University (Boston). by John Fiala... but appears to have since moved to Texas and is working under Kristen Harris.


  • Website/Ref: Jinx (Martone et al. 2008)
  • Cost: FREE
  • Platform: All (Java)
  • About: Tool for segmentation of 3D ET datasets. Many great features - loads NIF ontologies, nice smoothing opitons and icon bar, but Java does make it a bit slow.
  • Developed: NCMIR/UCSD (San Diego). Sole author: Steve Lamont


  • Website/Ref: bioView3D
  • Cost: FREE
  • Platform: All (Qt/OpenGL)
  • About: Open source volume renderer for biologists to visualize 3D ET, confocal, etc. Although still under development, can do some analysis and segmentation using a flexible/hierarchal XML structured file. NOTE: Not to be confused to with "BioVis3D", which is a (non-free) program.
  • Developed: Center for Bio-Image Informatics, UCSB (Santa Barbara). Primary author: Dmitry Fedorov


  • Website/Ref: Seg3D
  • Cost: FREE
  • Platform: Window or Mac (I believe)
  • About: Is a suite used to automatically segment volume in 3D - although mostly MRI data. The same group also did ImageVis3D - which is a fantastic volume renderer which works on the iPhone. Ricky recommended Seg3D to me, but haven't checked out yet.
  • Developed: NIH Center for Integrative Biomedical Computing at the University of Utah Scientific Computing and Imaging (SCI) Institute. Primary author: not sure yet!

Commercial Segmentation Programs

NOTE: Although I have listed prices for some of these products, keep in mind prices change over time. For most of these products the price isn't on their website - you have to e-mail to ask, and depending on your situation they may offer a discount.


  • Website/Ref: Amira homepage
  • Cost: $4,000 USD + $800/year support (2009)... although now has much more ugly/complex pricing structure where each feature is priced separately (eg: Amira Mesh Option $360). I believe at NCMIR we pay ~$9000/year for five user-license.
  • Platform: Windows 32 & 64 bit, Mac and Linux
  • About: Powerful visualizer (for both models and image files), plus filtering options, and good for making movies although animation process is quite tricky to master.


  • Website/Ref: Imaris (Bitplane AG, Zurich, Switzerland) (Haber et al. 2006 [4])
  • Cost: 16,000 USD! + 20% per year license (2009)
  • Platform: Windows only
  • About: Many similar features to Amira, but more limited. NOT worth the cost!


  • Website/Ref: BioVis3D homepage
  • Cost: 900 USD for single-computer-at-a-time licensee + 1 year e-mail support & minor upgrades
  • Platform: Windows only (and only 32-bit)
  • About: Does image alignment, segmentation and viewing in 3D. Segmentation appears to feature a great magnetic lasso tool (like photoshop) then drop the contours onto the next slice, resize and right click "fit contours to image" (see video). Newest version also includes functionality to draw tree structures to model neural pathways (see also: Neurolucida). Analysis looks very basic (volume, contour area stuff only).


  • Website/Ref: Osirix homepage
  • Cost: free for 32 bit version or $150 US for 64-bit version (2009)
  • Platform: OS X only
  • About: Used mostly to view cat scan / medical images in DICOM format (series of images in a folder). Has a nice "iTunes like" interface for listing "patients" and features 2D, 3D and 4D viewer with some nice presets for viewing bone, skin etc. Tends to crash sometimes though.


  • Website/Ref: AVS (Advanced Visual Systems Inc., Waltham, MA)
  • Cost: Unsure
  • Platform: -
  • About: Never tried, but think is more for programmers, not practical segmentation.


  • Website/Ref: Neurolucida homepage (MBF Bioscience Inc,) (Glaser and Glaser 2008[5])
  • Cost: Have to get quote, but based on my web searches is between $30,000 - $50,000!
  • Platform: Unsure
  • About: Videos very quite impressive - a tone of filtering options and some fairly nice segmentation tools, such as the "bifurcate" (drawing lines along dendrides in a tree-like structure), specifically for segmenting brain cells (cell body and dendrites). Most data is confocal LM, but also works on ET data. On their page they claim: Our flagship products Stereo Investigator and Neurolucida are the most widely-used analysis systems for stereology and neuron reconstruction. Note: MBF Biosciences was called "MicroBrightField" until a name change in 2004.

Web-Based / Server Side Collaborative Tools


  • Website/Ref: WIB (none yet)
  • Cost: FREE
  • Platform: any web browser with Flash plugin. See it working here.
  • About: A tool under development, but already very promising.
  • Developed: NCMIR, UCSD


  • Website/Ref: Viking (Anderson et al. 2010 [6])
  • Cost: FREE
  • Platform: Windows only (pretty sure)
  • About: The paper is terrific - describes a system specially developed for connectomics (neaural datasets) where user draws spheres and using drag-and-drop connects them together to show the topology of neural cells (bifurcations etc).... and similar process to draw synapses. After this process a brilliant graph can be generated showing what cells are connected to every other cell. The system has been tested using a 16.5 TB dataset of retinal connectome volume with three-six conncurrent users.
  • Developed: University of Utah, Main author is James Anderson


  • Website/Ref: CATMAID (Saalfeld et al. 2009 [7])
  • Cost: FREE
  • Platform: Any browser - see it working here.
  • About: The paper is a 3-page applications note well worth reading. In my opinion the system itself isn't as nice as WIB in that unloaded tiles are black.... and the only annotation is dropping text-labels (not segmenting). That said, images are all processed into a Zoomify format (optimized for speed) and they seem to have a very fast server too. Also, they have paired up with the Open Connectome Project which is well worth looking at.
  • Developed: Max Plank Institute.

Ssecrett and NeuroTrace

  • Website/Ref: Ssecrett (Jeong et al. 2010 [8])
  • Cost: Not publically available
  • Platform: Mostly cross platform - NeuroTrace is Qt, Ssecrett is C++but plans to go into Qt and merge into NeuroTrace interface.
  • About: Looks like a pretty brilliant solution to hybrid segmentation using a "shrinkwrap" approach + also tries to project this into ajacent slices along an axon.
  • Developed: Harvard University (Boston)

Single Particle Analysis

Although it uses an electron microscope, single particle analysis is very different from electron tomography, however the web-based program below is brilliant because it links together many programs in a single a web interface.


  • Website/Ref: Appion (Lander et al. 2010 [9])
  • Cost: FREE, but you need to contact SCRIPPS to become tester.
  • Platform: Should works in any browser (although underlying services maintained on many machines)
  • About: An integrated, database-driven pipeline to facilitate EM image processing. Features brilliant interface where user simply uploads their micro-graphs, then selects what programs to use then takes it through the full pipeline/workflow to reconstruct their particle.... at each step you can see various output including images and graphs.
  • Developed: SCRIPPS (San Diego)


Here I've listed some other resources on electron tomography:

  • - a terrific site which lists hundred of programs! This resource is designed to help the neuroscience community find the right tool for analysis, segmentation etc, and although most of these tools are targeted more at MRI, it's a great resource for electron tomography people too.
  • International Neuroinformatics Coordinating Facility (INCF) - a facility which wanting to introduce standards into the NeuroScience domain. They collaborate with and their website is hosted by NCMIR at the University of California, San Diego.