基于头部特征的人体检测

视觉监控是计算机视觉领域中近年来备受关注的前沿研究方向,视觉人体运动分析是视频监控中的研究重点。通过提取俯视图中人体头部区域特征,进行人体检测。由于人体头部区域的类圆形特征,一种圆形人体头部模型被提出,如何检测到合适的圆是人体检测的关键问题。本文采用对光照变化有较强鲁棒性的灰度梯度特征来检测人头。精确的实验表明了该算法在人体头部检测方面有较高的准确性和稳定性。

Human Body Detection and Tracking Based on Body-head Feature

The forefront research of the computer vision is the visual surveillance in recent years;especially the human motion analysis is focused on the study.This paper introduces a way of human body detection by extracting the regional characteristics of the human-head.We model the contour of human-head appearances by the circle as the circular feature,so how to detect the appropriate circle is the key to the problem.In this paper,we evaluate the model by the gray gradient feature which has strong robustness to li...