Texture Image Retrieval Using Non-separable Wavelets and Local Binary Patterns

This paper proposes a new image retrieval method using non-separable discrete wavelets (NDWT) and local binary patterns (LBP). Compared with the traditional wavelet, the three high frequency sub-images generated by the non-separable wavelets can extract more information and do not extensively focus on the three special directions any more. Further, local image texture and their occurrence histogram called local binary pattern (LBP) are a very powerful texture feature in spatial domain. Experimental results demonstrate the efficacy of the proposed algorithm in comparison with the existing related algorithms.