This research proposes a new mapping technique of Local Binary Patterns (LBPs) for texture classification. Mapping is an approach for producing a vector of features from the features that are extracted. This mapping method is based on extending nonuniform patterns for better classification of defects in patterned fabrics. By extending nonuniform patterns, a new mapping technique is suggested that extracts more discriminative features from textures. This new mapping can be used for various types of LBP and is tested for CLBP operator to show the improvement on the accuracy of the classification. The developed mapping technique is rotation invariant and has all the positive points of previous approaches. The proposed approach can code nonuniform patterns into more than one feature for producing distinctive features and better classification rate. Implementation of the proposed mapping on our patterned dataset shows that proposed method can improve the classification accuracy. Besides, the suggested approach improves the classification rate for all types of LBPs, particularly those with large neighborhoods.
