A New Logical Compact LBP Co-Occurrence Matrix for Texture Analysis

Texture is an important spatial feature, useful for identifying objects or regions of interest in an image. Statistical and structural approaches have extensively studied in the texture analysis and classif ication whereas little work has reported to integrate them. One of the most popular statistical methods used to measure the textural information of images is the grey-level co-occurrence matrix (GLCM). The present paper combines the Logical Compact LBP with OR operator (LCLBP-OR), which is derived on textons, w ith GLCM approach and LCLBPCM using three stages. The LCLBP-OR reduces the texture unit size from 0 to 255 to 0 to 15 and achieves much better rotation invariant classif ication than conventional LBP. The LCLBP-OR values are obtained by applying the logical OR operator in betw een relative positions of LBP w indow. To evaluate micro texture features in stage one textons are evaluated. To make texture features relatively invariant with respect to changes in illumination and image rotation LCLBP-OR images are applied on LBP images of texton shapes in stagetw o. Later in stage three the GLCM is constructed on LCLBP-OR and first and second order statistical features are evaluated for precise and accurate classif ication. The experimental results indicate the proposed LCLBPCM method classif ication performance is superior to that LBP, Gabor and other methods.