Performance evaluation of multi-target tracking using the OSPA metric

Performance evaluation of multi-target tracking algorithms is of great practical importance in the design and comparison of tracking systems. Recently a consistent metric for performance evaluation of multi-object filters (referred to as OSPA metric) has been proposed. In this paper we describe how the OSPA metric can be adapted to evaluate the performance of multi-target tracking algorithms. The main idea is to introduce the track label error into consideration in order to capture the data association performance, in addition to the existing cardinality and localisation errors. The paper demonstrates the proposed method by assessing and comparing the performance of two particle filters for multi-target tracking.