摘要
为了解决传统模板匹配算法在人脸检测中检测率低和速度慢的问题,提出一种新的人脸检测算法(BSICP)。引入最佳相似性作为相似性度量,只考虑匹配图像之间的相似点,减少错误匹配;以尺度迭代最近点算法作为搜索策略代替传统的逐点扫描匹配方法,加快检测速度。实验结果表明,该算法在IMDB-WIKI数据库中的五组变换图像下检测率均能达到97%以上,而且速度保持在0.076 s左右,具有很好的检测效果。
In order to solve the problem of low detection rate and slow speed of traditional template matching algorithm in face detection, a new face detection algorithm(BSICP) is proposed.The best similarity was introduced as the similarity measure, and only the similarity between matching images was considered to reduce the error matching.The scale-iterative nearest point algorithm was used as a search strategy instead of the traditional point-by-point scan matching method to accelerate the detection speed.Experimental results show that this algorithm can achieve a recognition rate of more than 97% under five groups of transformed images in the IMDB-WIKI database, and the speed remains around 0.076 s, which has a good detection effect.
作者
王钊
刘广瑞
孟少飞
Wang Zhao;Liu Guangrui;Meng Shaofei(School of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou 450001,Henan,China)
出处
《计算机应用与软件》
北大核心
2021年第11期215-218,共4页
Computer Applications and Software
