基于肤色及脸部特征的脸像检测及其应用

Face Recognition and its Application Based on Skin Color and Facial Features

第二代居民身份证的普及势在必行,由于办证数量多、质量要求高、时间安排紧以及为了方便群众,需要制定数字相片的相关规范标准。按照身份证制件要求,本文提出了一种将肤色和脸部特征相结合的方法来检测人脸及定位眼睛位置,并进一步检测及调整人脸色彩,从而根据检测结果来判定该证件相片是否符合制作标准。另外还根据要求对检测出的人脸部分在Lab色彩空间下进行色彩调整,具体步骤包括先对数字相片的颜色空间进行归一化调整以消除偏黄、偏绿和过亮或过暗的现象,再变换色彩空间为Lab空间,最后根据标准判断人脸部分的色彩是否符合要求并进行调整。实验证明该方法具有良好的检测效果。

Now the prevalence of the second edition ID card is imperative under the situation. For its numerous amounts, highly-quality demand and pressing arrangement, it is necessary to establish some criterion for the digital photo of ID card. According to the demand of ID card, this paper proposed a face detection and eye detection algorithm based on skin color and facial features, and then skin color adjustment. It also judges whether the digital photo accords with criterion or not by the results, Besides, color adjust- ment was made in Lab color space for the detected face region according to the demand. The main steps included unifying the color space to eliminate abnormal situations such as expesal, transforming to Lab color space, and color adjusting. The experiment results show that the method is effective