摘要: 提出一种基于AdaBoost的人脸性别分类方法,从一张低分辨率灰度人脸图像中辨认出一个人的性别。将启发式搜索算法融于AdaBoost算法框架中,从而发现新的可用于更好分类的特征。利用该方法进行人脸性别分类方面的实验,当使用少于500个像素比较时,正确识别率达到了93%以上,这与迄今已公布的最佳的分类器支持向量机(SVM)的正确识别率相当,但速度却快得多。
关键词:
性别分类,
AdaBoost,
启发式搜索
Abstract: This paper presents a method based on AdaBoost to identify the sex of a person from a low resolution grayscale picture of their frontal facial images. A heuristic search algorithm is used within the AdaBoost framework to find new features providing better classifiers. The experiments result of gender classification with the method presented in this paper indicate that the method is extremely fast and achieves over 93% accuracy with less than 500 pixel comparisons operations, these match the accuracies of the SVM-based classifiers which the best classifiers published to date.
Key words:
Gender classification,
Adaboost,
Heuristic search
朱文球;刘 强. 融合AdaBoost和启发式特征搜索的人脸性别分类[J]. 计算机工程, 2007, 33(02): 171-173.
ZHU Wenqiu; LIU Qiang. Syncretize AdaBoost Learning and Heuristic Search to Select Features for Gender Classification of Frontal Facial Images[J]. Computer Engineering, 2007, 33(02): 171-173.