Impact of topology-related attributes from local binary patterns on texture classification

A general texture description model is proposed, using topologyrelated attributes calculated from Local Binary Patterns (LBP). Theproposed framework extends and generalises existing LBP-based descriptorslike LBP-rotation invariant uniform patterns (LBPriu2), and LocalBinary Count (LBC). Like them, it allows contrast and rotation invariantimage description using more compact descriptors than classic LBP.However, its expressiveness, and then its discrimination capability, ishigher, since it includes additional information, including the number ofconnected components. The impact of the di erent attributes on textureclassi cation performance is assessed through a systematic comparativeevaluation, performed on three texture datasets. The results validate theinterest of the proposed approach, by showing that some combinationsof attributes outperform state-of-the-art LBP-based texture descriptors.