Design and analysis of an Event Indicator Function classifier for immune cell tracking applications

Recent advances in cell culture and cell imaging have made possible the automated acquisition of cell images. The automatic analysis of cells in such huge sets of images allows fundamentally new questions to be addressed in several biological fields such as immunology, proteomics, genomics and stem-cell research. Existing automated systems are not portable across a variety of cell videos because of random errors in both detection and tracking modules. These errors have to be identified and corrected to achieve tracker portability across videos. In this paper, we propose Event Indicator Function (EIF) classifier to predict detection and tracking errors in each frame using a set of features that are collected during tracking. It also predicts cell phenotypes (division and death), to accurately construct lineage tree (parent-daughter relationship) which has high significance in biological community. EIF classifier performance has been evaluated on variety of microscopic videos that differ both in cell density and dynamics. This approach helps in understanding the underlying system behavior and tracking can be improved using human assistance.