Color Context Binary Pattern Using Progressive Bit Correction for Image Classification

Local binary pattern (LBP) is sensitive to noise. Noise-resistant LBP (NRLBP) addresses this problem by thresholding local neighboring pixels into three-valued states (i.e., 0, 1 and uncertain bits) recovering bits via an error-correction mechanism. In paper, we extend NRLBP deal with color images propose a robust image descriptor called Color context (CCBP). CCBP, employ scale neighbor progressively correct the encoded bits. First, encode intra-channel in space use majority voting all across scales. Then, compute inter-channel feature distances bit propagation. Finally, construct concatenating histograms based on corrected codes. Experiments four benchmark databases demonstrate robustness of CCBP for classification under very low signal-to-noise ratio levels.