局部二值模式共生直方图的人脸识别

Face recognition based on local binary pattern histogram projection

传统局部二进制模式的人脸识别效率低,为此提出一种局部二值模式共生直方图的人脸识别算法。首先提取人脸图像的局部二进制模式共生直方图特征,然后采用最小二乘支持向量机建立人脸识别的分类器,最后采用多个人脸库进行有效性测试。结果表明,局部二值模式共生直方图的人脸识别算法的识别率达到95%以上,降低了计算的复杂度,单幅人脸识别时间明显少于对比算法。

The efficiency of face recognition is low, so a face recognition algorithm based on local binary pattern histogram projection is proposed. Firstly, local binary pattern histogram projection features of ace image are extracted, and then least square support vector machine is used build classifier of face recognition, finally the effectiveness is test- ed by some face database. The results showed that recognition rate of the proposed algorithm reaches more than 95% which reduces the computational complexity, and a single face recognition time was significantly less than that of the contrast algorithms.