SplineDist: Automated Cell Segmentation with Spline Curves
Encadrants : Vannary Meas-Yedid
Email : vmeasyed@pasteur.fr
Disponible : NON
Nombre d'étudiants : 1
Description :
Cell segmentation is a critical and challenging task in bioimage analysis field. Actually, cell detection is useful for counting, tracking or analyzing their morphology. Recently, several convolutional neural networks models have been proposed to tackle this task, especially the popular Stardist model. While StarDist describes cells as star- convex polygons, which is not always the case, SplineDist uses a more flexible and general representation by modelling objects as planar parametric spline curves.
Travail demandé : After the bibliographic study, the work consists in implementing SplineDist and test it on different modalities of images : fluorescent, phase contrast, immuno chemistry (color images) and compare its performances with the StarDist model already integrated in Icy (http://icy.bioimageanalysis.org/).
Liens complémentaires : 1. U. Schmidt, M. Weigert, C. Broaddus, and G. Myers, “Cell detection with star-convex polygons,” in Proceed- ings of MICCAI’18, Granada, Spain, September 16-20, 2018, pp. 265–273
2. Soham Mandal; Virginie Uhlmann, SplineDist: Automated Cell Segmentation with Spline Curves, ISBI, 2021 IEEE 18th International Symposium on Biomedical Imaging
2. Soham Mandal; Virginie Uhlmann, SplineDist: Automated Cell Segmentation with Spline Curves, ISBI, 2021 IEEE 18th International Symposium on Biomedical Imaging
