Tracking moving objects in video using enhanced mean shift and region-based motion field

In this paper, we propose a scheme for moving object tracking from videos by combining mean shift and motion field statistics. For mean shift, we employ an enhanced spatial-range mean shift that enables a reduced number of over-segmentation. For motion statistics, we combine the optical flow and high-order moment to generate motion regions that are associated with moving objects (or object parts). Experiments have been conducted on several indoor and outdoor (color/gray-scale) image sequences ranging from simple to median complexity. To evaluate the performance, three objective criteria are applied in addition to the visual inspection. The results show that the proposed method is promising for moving object tracking in video, with an averaging detection rate of 95%. Further, the proposed scheme is compared with that using the conventional mean shift for the tracking, indicating a significantly reduction in false alarm (≈ 30%).