【摘要】 该文运用人脸结构特点和数学形态学提出了一种稳健有效的人脸结构定位方法,该方法主要用于视频头肩序列图像的编码,对减小人脸预测误差、提高编码效率起着极为关键的作用。更多还原

【Abstract】 A robust and efBcient method of human face fine location is presented using human facial structure and mathematical morphology. The algorithm, which is mainly used in head-and- shoulder video coding, plays a critical role to decrease human face predictive discrepancy and increase coding efficiency.更多还原