摘要
Dalal等人提出的基于梯度方向直方图(HOG)的人体识别算法具有较高精度,因此成为了人体识别领域热门研究的对象,文中提出了一种快速HOG特征提取算法,并用支持向量机(SVM)训练得到人体分类器。实验结果表明,该算法在不降低检测效果的情况下,明显提高了人体检测的速度,且具有较强的鲁棒性。
Human detection based on HOG is a crucial issue in the field of human recognition. A fast HOG feature extraction algorithm is integrated, and SVM is used to train the human classifier. The experiments show that the algorithm does not reduce the detection effect, improved the speed of the human detection significantly, and has strong robustness.
出处
《电子设计工程》
2012年第11期190-192,共3页
Electronic Design Engineering
