基于外观模型的目标跟踪算法研究进展

基于视觉的目标跟踪是模式识别、计算机视觉、机器学习等多个学科的交叉研究课题,在视频监控、视频压缩编码、视频检索、智能交通等领域有着十分广泛的应用。为了使国内外同行对基于外观模型的目标跟踪方法有一个较为全面的了解,对其进行了系统总结。在介绍跟踪算法原理的基础上,重点阐述了两大类基于外观模型的目标跟踪方法:产生式方法和判别式方法,深入讨论了其中的典型算法和研究成果,并对这些算法在公开数据集上的测试结果进行了分析比较,最后展望了该领域未来的发展方向。

Object tracking algorithms based on appearance models:A survey

Visual object tracking is a hot topic in areas of pattern recognition, computer vision and machine learning, which has a wide range of applications in video surveillance, video compressing coding, video retrieval, intelligent transport system and so on. In order to make a better understanding of object tracking algorithms based on appearance models for domestic and international peers, we comprehensively review the latest research progress in this field. We firstly introduce the mechanism of tracking algorithms.Then we summarize current researches on visual object tracking based on appearance models, including generative and discriminative methods. Thirdly, the state-of-the-art tracking algorithms are compared on the public datasets. We finally discuss future research trend of object tracking.