摘要
针对人脸图像性别识别中单一特征识别率不高的问题,提出了融合纹理特征和形状特征的人脸图像性别识别方法.通过局部二值模式(LBP)及其改进算法提取人脸图像的纹理特征,梯度直方图(HOG)提取人脸图像的形状特征,融合两个特征利用Adaboost分类器进行人脸图像的性别分类.在ORL人脸数据库和自制人脸数据库CZB上的实验结果表明,相对于直接利用像素特征和单一特征的识别方法,融合多特征的人脸性别识别方法的识别率明显提高.
In order to improve low recognition rate by using single feature in face image gender recognition,a face image gender recognition method based on texture feature and shape feature is proposed. Local binary pattern(LBP) and its improved algorithm are used to extract texture feature,Histogram of oriented gradient(HOG) is used to extract shape features of face images. Two features are fused to classify the gender of face images using Adaboost classifier. Experimental results on ORL face database and CZB face database show that the recognition rate of face gender recognition method based on multi-feature fusion is significantly higher than that based on pixel feature and single feature directly.
引文
[1]Saatci Y,Town C.Cascaded classification of gender and facial expression using active appearance models[C].IEEE International Conference on Automatic Face and Gesture Recognition,2006:393-398.
[2]Shan C.Leaning local binary patterns for gender classifications on real-world face images[J].Pattern Recognition Letter,2012,33(4):431-437.
[3]孙宁,冀贞海.基于局部二元模式算子的人脸性别分类方法[J].华中科技大学学报,2017,35(3):177-181.
[4]郑瑶娜.基于人脸图像的性别识别和年龄估计[D].合肥:合肥工业大学,2017.
[5]Wang J G,Li J,Chong Y L,et al.Boosting dense SIFT descriptors and shape contexts of face image for gender recognition[C].IEEEcomputer society conference on Computer Vision and Pattern Recognition,2010:96-102.
[6]Wang J G.dense SIFT gabor descriptors-based face representation with applications to gender recognition[C].International conference on Control Automation Roberts&Vision.2010:1860-1864.
[7]裴子龙.基于卷积神经网络的人脸图像性别识别研究[D].临汾:山西师范大学,2017.
[8]Qin H,Yan J,Li X,et al.Joint training of eascaded CNN for face detection[C].Proceedings of the IEEE conference on Computer Vision and Pattern Recognition,2016:3456-3465.
[9]李森林,石元泉.一种基于SAE和BP网络相结合的人脸识别模型[J].怀化学院学报,2017,36(5):78-82.
[10]张铮,徐超.数字图像处理与机器视觉(第二版)[M].北京:人民邮电出版社,2014:353-354.
[11]Dalai N,Tfiggs B.Histograms of oriented gradients for human detection[C].Proceedings of IEEE computer society conference on Computer Vision and Pattern Recognition,2005:886-893.
[12]陈松.基于肤色的人脸检测与识别[D].成都:电子科技大学,2007.