Multi-target Tracking by Discriminative Analysis on

This paper addresses the problem of multi-target tracking in crowded scenes from a single camera. We propose an algo-rithm for learning discriminative appearance models for dif-ferent targets. These appearance models are based on covari-ance descriptor extracted from tracklets given by a short-term tracking algorithm. Short-term tracking relies on object de-scriptors tuned by a controller which copes with context vari-ation over time. We link tracklets by using discriminative analysis on a Riemannian manifold. Our evaluation shows that by applying this discriminative analysis, we can reduce false alarms and identity switches, not only for tracking in a single camera but also for matching object appearances be-tween non-overlapping cameras. Index Terms — tracking, controller, re-identification, co-variance matri