把形变视频跟踪看作一个动态图像配准问题,基于已有的模糊规则自适应学习弹性图像配准方法,提出了一个具有较好鲁棒性的视频跟踪算法。较之于一些经典的跟踪算法,提出的新算法显示出了更鲁棒的跟踪效果。
By taking deformal visual tracking as the procedure of active image registration,the existing fuzzy-inference-rule adaptive elastic image registraion is used to develop a much more robust visual tracking algorithm.Compared with two classical visual tracking algorithms,the proposed algorithm demonstrates a promising performance.