摘要
在彩色图像的人脸检测中,针对人脸分类器训练时可能存在的过拟合问题,提出了一种基于肤色与改进Adaboost算法的人脸检测。双阈值的权重更新方式避免了分类器训练过程中可能出现的过拟合现象;人脸检测时利用肤色的聚类性,将检测到的肤色区域作为Adaboost分类器的搜索候选区域。实验证明:在复杂背景中,提出的方法有效地缩短了检测时间,降低了误检率。
Aiming at potential over-fitting problem in face classifier training in face detection of color image,a novel method based on skin color and improved Adaboost algorithm is proposed. Weight updating mode of doublethreshold is employed to avoid the over-fitting phenomenon that may occur in the classifier training. In face detection,clustering performance of skin color is used,detected skin color area are used as search candidate regions of Adaboost classifier. The experimental results show that the proposed method can effectively shorten the detection time and reduce the false detection rate in complex background.
作者
沈翔
朱建鸿
SHEN Xiang;ZHU Jian-hong(Key Laboratory of Advanced Process Control for Light Industry,Ministry of Education,Jiangnan University,Wuxi 214122,China)
出处
《传感器与微系统》
CSCD
2019年第4期143-145,共3页
Transducer and Microsystem Technologies