摘要
步态识别中大多采用步态轮廓作为识别特征,因此提取完整封闭的运动人体轮廓以准确表达步态特征是正确识别的前提。本文提出一种采用高斯模型的步态轮廓分割算法。在人的运动方向与摄像机成像面平行和摄像机静止的条件下,假设序列图像所有帧中对应像素点背景时刻的灰度值在时间轴上是高斯分布,而目标时刻不满足这种分布,采用统计推断的方法分割出运动目标轮廓。实验结果表明,本文算法不仅能够提取出完整的人体轮廓,并且能有效地去除噪声,对阴影抑制也有一定效果,能够提高步态识别率。算法直接在RGB空间或灰度空间进行,无需进行颜色空间转换,也无需建立单独的背景图像,计算量小,处理实时性高。
Gait silhouette is mostly used as recognition feature in human gait recognition, so the prerequisite of correct recognition is to extract complete and closed moving human silhouette to express gait feature ex- actly. A gait silhouette extraction algorithm using Gauss model is proposed. Under the condition that the motion direction of object is parallel to the imaging plane of the camera and the camera is static, it is sup- posed that in the serial images, the distribution of the pixel＇s gray level at the time of background on time axis is Gaussian distribution, while the distribution at the time of object does not satisfy this distribution, then moving object silhotlette is extracted using statistical inference approach. The experimental results demonstrate that the proposed algorithm can not only extract human silhouette completely and accurately, but also can remove noise and shadow and increase recognition rate. The algorithm can be directly used on RGB space or grey space, without changing color space or building the background model, it is less computative and has high real-time processing performance.
出处
《传感技术学报》
CAS
CSCD
北大核心
2008年第7期1155-1159,共5页
Chinese Journal of Sensors and Actuators
基金
重庆市自然科学基金资助项目(CSTC2006BB2155)
关键词
步态识别
轮廓提取
高斯模型
运动目标分割
gait-based recognition
silhouette extraction
gauss model
moving object segmentation
