Combining gray-level co-occurrence matrix and statistics features for rotation invariant texture classification in wavelet domain

In order to improve the accuracy and efficiency of rotation invariant texture classification, we develop a novel classification method based on gray-level co-occurrence matrix and discrete wavelet transform in the paper. The method combines the probability of specific neighboring resolution cell pairs occurred at the same time in the whole image and statistics features in wavelet domain. Discrete wavelet transform is firstly adopted to decompose images into several sub-bands. Then probability and statistics features are extracted from these different sub-bands. The probability is calculated from the approximation sub-band and statistics features are calculated from both approximation sub-band and detail sub-bands. Finally, the method combines the probability and statistics together as features for rotation invariant texture classification. Experiments are conducted on two texture image sets and the results of experiments show the good performance of our method.