关键词: 计算机视觉, 人体姿态, 深度图像, 二元纯位相匹配滤波器（BPOF）, 随机森林

Abstract: Human pose estimation is a research topic which has wide application prospect, and the topic has become one of the important research hotspot in the field of computer vision. For feature extraction in human pose estimation, the paper proposes an improved method based on Binary Phase-Only Filter（BPOF） algorithm, first of all, it calculates the scanning line length value from eight directions of every pixel in the image, then puts the eight value into the BPOF algorithm to get the feature of the pixel, it also optimizes the random forest which is used to classify, finally can estimate the human pose. The improved method has made a very big enhancement in recognition rate and robustness, at the same time, the optimized random forest classifier reduces the algorithm system time consumption.

Key words: computer vision, human pose, depth image, Binary Phase-Only Filter（BPOF）, random forest