视频流分析的隐藏马尔可夫链算法

视频流分析是多媒体技术研究的重要内容，具有广泛的应用领域．应用马尔可夫链模型，研究视频中的人体运动目标位置的跟踪，通过对跟踪的第一帧视频中的人体区域进行手工标注，建立马尔可夫链模型．然后结合卡尔曼估计算法，跟踪出下一帧的人体位置．实验证明，这种算法可以有效地解决在复杂的运动背景下，人体运动目标的跟踪问题．

Applying Hidden Markov Chain Algorithm to Video-Stream Analysis

Tracking a moving target is the main topic in video stream analysis which is a hotspot in the study of multimedia technology. In this paper, the hidden Markov chain algorithm is applied to tracking moving target position of human body in motion video. A Markov chain model is built at first by marking manually the area of human body to be tracked at the first frame of video. And then, the position of human body at the next frame of video is tracked down by combining with kalman algorithm of estimate. As proved by experiment, this algorithm is able to settle effectively the problem of tracking moving target of human body against a complex moving background.