Model&Motion based Shape Tracking in Large-Scale Cellular Datasets

Sha Yu, Yao Lu, Derek Molloy

Abstract: This material proposes a model&motion based new cell tracking framework, where two main tasks are targeted, including shape segmentation in low contrast images, and estimation of cell trajectories in clutter environment. To be concrete, the parametric snake model and the optic-flow technique are seamlessly combined to cooperate with each other. By that means, the shortcomings of each side are addressed in the process of simultaneously tracking cell deformation and movement. By experiments on real cellular datasets, the proposed approach has been demonstrated to be well applicable for jointly segmenting and tracking cells, particularly under multiple challenges, such as low quality images, ambiguous cell boundaries, and closely touching cells in high concentration.

External IDs:dblp:conf/icmlc2/YuLM18