Combination of DWT and Co-occurrence Matrix Features for Human Emotions Description

In this paper, we present an approach for human emotion description based on proposed algorithm combines the strengths of two methods of analyzing textures, the Discrete Wavelet Transform (DWT) and the Grey Level Co-occurrence Matrix (GLCM). The proposed algorithm is introduced to extract the texture features by combining (1) wavelet statistical features extracted from the approximation and detail regions of three level discrete wavelet transform decomposed images and (2) grey level co-occurrence matrix features extracted from original image, approximation and detail sub-bands of one level discrete wavelet transform decomposed images at different angles . Experimental results are presented using image database consists of a set of photograph face images describing some human emotions, these images were generated by using a digital video camera. The experimental results show that the proposed algorithm is effective and achieves high similarity ratios.