表情识别

  在频域中组合空间域和局部相位量化（LPQ）的局部二进制模式（LBP）用于面部识别。 首先，从LBP特征和频域中提取人脸图像以提取LPQ特征，然后合并到LBP / LPQ直方图中以比较直方图相似度。 最后，使用最近邻原则来识别。 在YALE和AR标准人脸数据库上的实验表明，该方法获得的结果优于单一方法。(A local binary pattern (LBP) combining spatial domain and local phase quantization (LPQ) in frequency domain for face recognition. Firstly, the face image is extracted from the LBP feature and the frequency domain to extract the LPQ feature, and then merged into the LBP /LPQ histogram to compare the histogram similarity. Finally, the nearest neighbor principle is used to identify. Experiments on the YALE and AR standard face databases show that the results obtained by this method are better than the single method.)