摘要
提出了结合图像灰度梯度信息和彩色直方图信息的人头识别算法.该算法中:灰度梯度信息可以排除背景图像中的颜色干扰,彩色直方图信息可以适应识别目标的形状变化和背景中的纹理对识别目标的影响.对2种算法的结果进行规一化处理,实现两者的相互补充.最后通过实验证明了该算法的有效性.
An algorithm that integrates gray scale gradier, t cue and color histogram cue to detect the human head is proposed. In this algorithm, gray scale cue is capable eliminating background color interferences, while the color histogram is capable of adapting the changing of shape of detecting candidates and the effect of texture in the background. Formalizing the result of two parts enabled the final conclusion tobe orthogonal to each other. Filmily, experiments applied prove the availability of the algorithm.
出处
《重庆工学院学报》
2007年第23期1-5,共5页
Journal of Chongqing Institute of Technology
关键词
灰度梯度
彩色直方图
人头识别
gray scale gradient
color histogram
head detection
