Multi-object tracking with inter-feedback between detection and tracking

Multi-object tracking is an important but challenging task in computer vision. Tremendous investigations have been made on the topics, among which tracking-by-detection method first detects objects independently at each frame and then links the detected objects into trajectories. One shortcoming of this method, however, lies in the fact that it regards detecting and tracking as two separated processes and the tracking information are not used in detection, which often results in many false and missing detections and involves heavy computational complexity. In order to solve this problem, this paper proposes a multi-type multi-object tracking algorithm, by introducing on-line inter-feedback information between the detection and tracking processes into the tracking-by-detection method. Our tracking algorithm consists of two iterative components: detection by feedback from tracking and Tracking based on detection. In the detection step, objects are detected by the detectors adjusted by information from tracking. In the tracking step, we use group tracking strategy based on detection. Moreover, in order to handle tracking scenarios with different complexity, objects are classified into two categories, i.e. single object and multiple ones, and are dealt with different strategies. The proposed algorithm is evaluated on several real surveillance videos and achieve higher performance in contrast to the state-of-the-art methods. Besides the high precision, it also has demonstrated that the proposed algorithm needs less detector and searching scale and can run in real time for many tracking applications.