摘要
对拍摄得到的驾驶员视频帧图像,使用复合肤色模型检测人脸;通过自适应边缘检测、图像增强等方法处理得到特征图像,经特征区域筛选,依据人脸先验知识匹配得到最佳人眼对;提取眼部特征向量,结合LVQ神经网络进行模式识别检测眼部状态,为判断驾驶员是否处于疲劳状态提供判据。
Capture color video images of driver,at first locate driver's face by mixing skin model ,second analyzed edge and gray information to search possible eye regions,using experiential knowledge to pick up a pair of characteristic regions which are best match.Finally distill eye's characters vector,which are processed by LVQ network to determine eye states.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第34期211-213,共3页
Computer Engineering and Applications