In this paper, a new feature extraction method using multiresolution local pattern information of log-polar autocorrelation image is proposed for robust image classification and retrieval. The robust image classification and retrieval systems are required when the query images are geometrically distorted. The present paper advances our recent work to obtain high performance of retrieval accuracy for image retrieval. In the simulations, the method is experimentally applied to characterization of texture images robust to geometrical (rotation, scaling and translation) distortion. It is efficiently used for texture retrieval system to demonstrate the usefulness of the method. It is shown that the higher retrieval rate is achieved compared with the conventional approaches