Volumetric Image Segmentation

Encadrants : Vannary Meas-Yedid
Disponible : OUI

Spécialité : IMA
Nombre d'étudiants : 1
Description :

Image segmentation is the first step in bio-image analysis, to extract the regions of interested and their relationships that can be highlighted and measured for further interpretation. Microscopies play a key role in the identification and characterization of cellular compartments, pathogens and host – pathogens interactions. The electron microscopy provides ultrastructural resolution of large volume images.

Pré-requis : Python
Travail demandé : Deep Learning models have achieved the state of the art in image segmentation. Several models have been proposed to segment 3D microscopic objects [1, 2, 3]. The work is to make a survey of different DL models to segment 3D organelles. Then one model will be chosen to segment 3D organelles. For the tests, we will leverage the open organelle site of Janelia datasets (https://openorganelle.janelia.org/) to perform the model.


1-Çiçek, O, Soeren L,Brox, Ronneberger, T: 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation. In: Ourselin, Sebastien, Joskowicz, Leo, Sabuncu, Mert R.., Unal, Gozde, Wells, William (eds.) MICCAI 2016. LNCS, vol. 9901, pp. 424–432. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46723-8 49
2- T. Sato and K. Hotta, "EncapNet-3D and U-EncapNet for Cell Segmentation," 2019 Digital Image Computing: Techniques and Applications (DICTA), 2019, pp. 1-7, doi: 10.1109/DICTA47822.2019.8945839.
3- Nguyen, T., Hua, BS., Le, N. (2021). 3D-UCaps: 3D Capsules Unet for Volumetric Image Segmentation. In: , et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2021. MICCAI 2021. Lecture Notes in Computer Science(), vol 12901. Springer, Cham. https://doi.org/10.1007/978-3-030-87193-2_52
DOI: https://doi.org/10.1007/978-3-030-87193-2_52