A Probabilistic Relaxation Labeling (PRL) Based Method for C. elegans Cell Tracking in Microscopic Image Sequences

摘要

Automatic cell tracking for 3D time-lapse images plays an important role in studying of live cell, which allows investigation of biological processes in vivo at the single cell resolution. In this paper, we propose a method for cell tracking during C. elegans embryogenesis based on probabilistic relaxation labeling (PRL). Instead of relying on the absolute cell position, we make use of the cell-to-cell relative position with two compatibility coefficients to achieve cell tracking. The tracking results obtained with this method are highly accurate with temporal resolution of one minute. In addition, our method can also been realized on multi-core CPUs, therefore providing an effective tool for analysis of large-scale data consisting of 3D time-lapse live cell images.