High‐speed visual tracking with mixed rotation invariant description

Abstract

A mixed rotation invariant description (MRID)-based tracking algorithm and a novel high-speed visual tracking system that implements the algorithm are proposed. MRID is a novel rotation invariant description of texture and edge information by annular histograms and dominant direction. It overcomes rotation variant and large computation issues in conventional local binary pattern histograms of oriented gradient (LBP-HOG) feature description. The proposed tracking system contains an image sensor, a hierarchical vision processor and a two dimension of freedom actuator. The vision processor integrates processors with pixel and row-level parallelism to speed up the tracking algorithm. Experiment results show that the proposed system can achieve over 1000 fps processing speed of the tracking algorithm.