摘要
为了提高复杂背景下多人脸检测率以及人脸检测速度,提出了一种基于改进AdaBoost、肤色检测和二维主成分分析法(Two-dimensional Principal Component Analysis,2DPCA)的人脸检测方法。该方法首先利用金字塔结构快速检测人脸,得到人脸检测区域,然后利用肤色检测对待判人脸区域进行过滤,过滤误检的非人脸区域,最后根据人脸的几何位置进行人脸关键部位的2DPCA检测。仿真结果表明,该方法实现了复杂背景下多人脸图像快速检测和精确定位,有效降低了误检率,使检测结果更加精确。
In order to improve the speed of face detection and face detection rate under complex background, proposed a face detection method based on the improved AdaBoost, skin color detection and 2 DPCA. First ,the method using the pyramid structure of rapid detection of human face, obtain the face detection area. What is more, using skin color detection to face region to filter, filter mistakenly identified the face region. Finally, detect the key parts of the face according to face geometric position. The simulation results show that the algorithm has realized the rapid detection and precise positioning of the more human face images under complex background, effectively reduce the detection error rate, make the results more accurate.
出处
《电子设计工程》
2014年第8期116-119,共4页
Electronic Design Engineering
