Section: New Results

3D Modeling of developing organisms

Participants : Gaël Michelin, Grégoire Malandain, Léo Guignard [Virtual Plants] , Christophe Godin [Virtual Plants] .

This work is made in collaboration with Patrick Lemaire (CRBM).

Image-based studies of developing organs or embryos produce a huge quantity of data. To handle such high-throughput experimental protocols, automated computer-assisted methods are highly desirable. We aim at designing an efficient cell segmentation method from microscopic images. Similary to another work  [32] , the proposed approach is twofold: first, cell membranes are enhanced or extracted by the means of structure-based filters, and then perceptual grouping (i.e. tensor voting) allows to correct for segmentation gaps (see figure 6 ). We assessed different structure-based filters as well as different perceptual grouping strategies to identify the most efficient combination, in term of result quality and computational cost [13] .

Figure 6. Illustrations of the different steps of the algorithm: (A) a 2D slice of original image, (B) the resulting surface detector response, (C) the directional extrema of the response image, (D) the deduced binarisation of the cell membranes, (E) the result of Tensor Voting applied to binarised image, (F) the cells segmentation computed from (E), (G) a 3D view of original image, (H) a 3D view of the cells segmentation.
IMG/fused_t090_161_sample.png IMG/kr_t090_161_4_sample.png IMG/kr_ext_t090_161_4_sample.png IMG/kr_bin_t090_161_4_sample.png
(A) (B) (C) (D)
IMG/TV10_090_fr_bin_5_161_sample.png IMG/wats_kr_bin_4_090-filtered_161.png IMG/fused_090_3Dview.png IMG/Movie_WAT_t000_to_t098-1-090.png
(E) (F) (G) (H)