多尺度中心误差补偿二值模式的表情识别-计算机工程与应用.pdf

148 2016 ，52（8） Computer Engineering and Applications 计算机工程与应用 多尺度中心误差补偿二值模式的表情识别 史久根，司小龙 SHI Jiugen, SI Xiaolong 合肥工业大学 计算机与信息学院，合肥 230009 School of Computer and Information, Hefei University of Technology, Hefei 230009, China SHI Jiugen, SI Xiaolong. Facial expression recognition based on multi-scale center error compensation binary pat- tern. Computer Engineering and Applications, 2016, 52 （8 ）：148-152. Abstract ：In order to solve the limitations of Gabor wavelet and Local Binary Pattern （LBP ）in facial expression recognition, a method based on multi-scale Center Error Compensation Binary Pattern （CECBP ）is proposed. In the method, the images are preprocessed and multi-scale pyramid of these images are then created. Center Error Compensation Binary Pattern （CECBP ）is utilized to encode the images of every layer in image pyramid. The chunked and encoded histogram sequences are used as a feature and Support Vector Machine （SVM ）is used in classification. To demonstrate the superiority of the proposed method over traditional Gabor wavelet and LBP, cross-validations on JAFFE, Cohn-Kanade and Pain expression database, it shows that the method can not only suppress noise, but also have state-of-the-art classification accuracy and faster recognition speed. Key words ：Gabor wavelet ；Local Binary Pattern （LBP ）; Center Error Compensation Binary Pattern （CECBP ）; Support Vector Machine （SVM ）; expression recognition 摘 要：针对Gabor 小波与局部二值模式（Local Binary Pattern ，LBP ）在表情识别上的局限性，提出了一种多尺度中 心误差补偿二值模式（Center Error Compensation Binary Pattern ，CECBP ）的表情识别方法。对预处理后的人脸表 情图像创建多尺度的金字塔，用中心误差补偿二值模式对金字塔中的各层图像进行编码，分块提取各层编码后的直 方图序列作为特征，用支持向量机（Support Vector Machine ，SVM ）进行分类。在JAFFE 、Cohn-Kanade 以及Pain Expression 表情库上的交叉验证表明，该方法可以抑制噪声，具有较高的识别率和较快的识别速度，比传统的Gabor 小波以及LBP 更具有优势。 关键词：Gabor 小波；局部二值模式；中心误差补偿二值模式；支持向量机；表情识别 文献标志码：A 中图分类号：TP391 doi ：10.3778/j.issn. 1002-8