3DMMS

This program implements the 3DMMS algorithm proposed in 3DMMS: Robust 3D Membrane Morphological Segmentation of C.elegans Embryo, Jianfeng Cao, Ming-Kin Wong, Zhongying Zhao, Hong Yan.

Usage

• Platform dependency

  This program is developed in Matlab 2017b. For computational efficiency, parallel computing is adopted in 3DMMS.

• Structure of folders

  3DMMS is used to segment C.elegans embryo imaged with fluorescent microscopy
  3DMMS/
    |--data/: nucleus information file from AceTree and raw membrane images
      |--aceNuc/
        |--170704plc1p2/: nucleus location file (refer to AceTree)
      |--colorMap.mat: color map used for labelling result (*_rgb.tif)
      |--membrane/170704plc1p2/: raw membrane image 
    |--Evaluation/: data for experimental evaluation in BMC Bio 3DMMS paper
    |--getNucFromacetree/: codes for get nucleus location in *.csv file
      |--transformed/: index map between label <--> cell name
      |--***
    |--MINS-1.3/: external library for fast computation
    |--results/:
      |--resultWithMerge/: results with two daughter cells possiblely merged.
        |--informationForMerge/: intermediate results in parallel computation
        |--merged_memebraen/: *.mat is prefered. *.tif is used for visualization.
        |--mergedResults/: stastial results of the final segmentations.
    |--TiffSave/: library for saving tif images.
    |--tinevez-matlab-tree-3d13d15/: library for constructing cell division tree.
    |--**      
  

• Parameters settting

  If you want to analyze your own dataset, you might need to change parameters according to your own dataset. I3DMMS.m function includes all the parameters you need to tune. They are listed as follows:

  Parameter name	Meaning
  data_name	membrane image dataset name
  max_Time	maximal time point of the embryo stack
  prescale	downsample ratio on each slice
  reduceRatio	downsample ratio on the whole embryo
  xy_resolution	resolution on each slice
  z_resolution	distance of each slice

• Experiment

  Suppse 3DMMS as the root folder,

  1. run startup.m first to generate colorMap.mat and add root path;
  2. Then run I3DMMS.m to get results as that in results folder.

• Note

  1. ImageJ can be used to view *.tif files, Plugins-->Process-->Show Color Surfaces.
  2. Nucleus location file CD**.csv is necessary which can be obtained by using AceTree and StarryNite.
  3. 3DMMS aims to segment cells in 3D, where cells closely contact to each other. Raw membrane stack images can come from C. elegans, Arabidopsis thaliana and Drosophila, but not individual cells, like cells in the blood.
  4. The dataset is not included in this repository. To run the code with the datset, you can either download the dataset from figshare or download the zip file(recommended).

Correspondence

Please contact to [email protected]