Degree of nonlinearity (DoN) measure for target tracking in videos

Performance evaluation of tracking methods includes methods of relative and absolute performance. Absolute tracking performance is the robust end result presented to a user which determines the product solution for real world analysis. However, to achieve robust performance, the tracking method is subject to the sensor data, filtering performance, and associated models, which requires relative performance evaluation. In this paper, we highlight the efficacy of using DoN measure in evaluating video tracking capabilities. Three developments are presented as to the real-world issues associated with the nonlinear video-based tracking: (1) challenges of performance evaluation with real data, (2) approaches to utilize DoN as improvements for relative track evaluation, and (3) operational implementations lessons associated from user-defined operating picture (UDOP) plugins. Results are presented using relevant data with a highlighted tracker.