Summary: | 碩士 === 中原大學 === 資訊工程研究所 === 101 === With the development of multimedia technology, video tracking applications are also emerging in people’s lives. Video tracking is often subject to the interfering factors, e.g., camera shake, object rotation, object deformation and occlusion, etc., leading to tracking failure. Therefore, this paper proposes a region-based tracking algorithm that uses the color region segmentation and matching, overlap processing of region distribution map, and set update (e.g., union, intersection, etc.) of regions, thereby improving the video tracking accuracy. In order to verify the effectiveness of our method, a video database with various video behaviors was established. The results demonstrated that our method could achieve the average tracking accuracy rate of more than 85%. The proposed method could be integrated with “feature-based video tracking” or “part-based recognition” to achieve more accurate tracking, and could be further applied for early childhood education.
|