Learning Context Cues for Synapse Segmentation

This code implements a synapse segmentation algorithm, as introduced in [1]. This is based on the first version available at http://cvlab.epfl.ch/software/synapse

Please check http://cvlab.epfl.ch/software/synapse for updated instructions on how to use this code.

xx WARNING xx: THIS IS A BETA VERSION and DOES NOT FULLY IMPLEMENT THE METHOD IN [1].
xx WARNING xx: If you want to run the approach of [1], download the original version
xx WARNING xx: from http://cvlab.epfl.ch/software/synapse

GETTING STARTED

You need CMake to compile the code. So far, only Linux is supported.

First create a build folder, for example:

  cd <where_iiboost_is>
  mkdir build

Now configure with ccmake:

  ccmake ../

Choose to build the Python Wrapper (BUILD_PYTHON_WRAPPER set to ON), specify the path to ITK 4.5 (ITK_DIR) and set CMAKE_BUILD_TYPE to RELEASE.

(Make sure ITK was compiled with -fPIC in CXXFLAGS)

Finally, if you are using ILASTIK, set PYTHON_BASE_PATH to the path where ilastik is. Otherwise, point it to your root usr folder (eg /usr).

You can now compile with make, and try the python wrapper with

  python2 python_test_class.py

REFERENCES

For more information about the synapse segmentation algorithm, please check the following article:

[1] Learning Context Cues for Synapse Segmentation C. Becker, K. Ali, G. Knott and P. Fua. IEEE Transactions on Medical Imaging (TMI) 2013

[2] Learning Context Cues for Synapse Segmentation in EM Volumes. C. J. Becker, K. Ali, G. Knott and P. Fua. International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), Lecture Notes in Computer Science, 2012.

CONTACT

Please mail [email protected] for bug reports, comments and questions.