Team, Visitors, External Collaborators
Overall Objectives
Research Program
Highlights of the Year
New Software and Platforms
New Results
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
XML PDF e-pub
PDF e-Pub

Section: New Results

Cytoplasm segmentation from cells confocal microscopy images

Participants : Somia Rahmoun, Eric Debreuve, Xavier Descombes, Fabienne de Graeve.

As part of the ANR project RNAGRIMP, two series of images have been acquired using fluorescence microscopy: one where the cell cytoplasm has been stained with GFP (Green Fluorescent Protein), the second where the nuclei have been stained with DAPI (4’,6-diamidino-2-phenylindole). The first steps are detecting the nuclei on the DAPI images and learning a classification procedure into living cell or dead cell based on morphological and radiometric nuclei properties (average intensity, area, granularity, circularity, ...).

The next step is to segment (i.e., extract automatically the region of) the cell cytoplasms on the GFP images. Indeed, the target RNP-IMP granules appear in that compartment of the cell and are visible through their GFP response. This segmentation problem is particularly difficult due the heterogeneity of the cells intensity. This heterogeneity even appears within a given cell. Besides, cells sometimes form cluster in which there is no clear separation between adjacent cells.

In this context, we have considered a two steps algorithm to segment the cytoplasm. The first step consists of the image segmentation in small areas called superpixels that represent adjacent pixels with similar intensity. An automatic algorithm based on the watershed transform has been chosen after evaluating and comparing different strategies (based on iterative clustering, minimum spanning tree, persistent edge selection ...).

The second step of the proposed approach performs superpixels merging to obtain the final segmentation. Starting from the previously detected nuclei to define cell seeds, the neighboring superpixels are merged iteratively if a radiometric similarity is detected. Ambiguities between neighboring cells are solved by combining radiometric and shape criteria. This cell growth process is considered layer by layer and performed in parallel.