lsm3d - Tools for segmenting 3D images of plant tissues at multiple scales using the level set method¶

One of the main technical challenges in the reconstruction of plant tissues from 3D confocal images remains the faithful segmentation at the cellular level. A typical issue with segmentation based on watershed methods is the weaker staining of some outer periclinal walls. This systematically causes segmentation errors as holes on the tissue surface.

We detect the outer surface of the tissue using level set method (lsm_contour). We enhance the signal of the outer periclinal walls by the detected outer surface. Also, we propose a new edge function, based on second order derivatives of the image, for a level set used to segment the tissue at cellular level (lsm_cells).

The "lsm3d" tools¶

- lsm_contour --> detects the outer surface of the tissue
- lsm_cells --> used to cellular segmentation or nuclei detection

Dependencies¶

The levelset tools make use of the CImg image processing C++ library, which in turn needs Xlib library. Furthermore, the lsm_cells tool is parallelized using OpenMP.

In particular, if you do not have Xlib on your computer, you may add it:
• on Ubuntu:
sudo apt-get install libx11-dev

• on Mac OS:
install xQuartz from www.xquartz.org

In a terminal go to the directory where you wish to download and install the tool

cd Path_to_your_directory

svn checkout http://forge.cbp.ens-lyon.fr/svn/lsm3d


Compile¶

cd lsm3d
./lsm3d_compile.sh


export PATH=\$PATH:Path_to_your_directory/lsm3d/bin