Survey on LBP based texture descriptors for image classification

摘要: The aim of this work is to find the best way for describing a given texture using local binary pattern (LBP) based approach. First several different approaches are compared, then fusion approach tested on datasets and compared with proposed in literature (for fair comparisons, when possible we have used code shared by original authors).Our experiments show that uniform quinary (LQP) rotation invariant pattern, where bin selection variance performed Neighborhood Preserving Embedding (NPE) feature transform applied, obtains method performs well all datasets.As classifier, stand-alone support vector machine (SVM) random subspace ensemble SVM. We compare descriptors our coupled outperforms other recent state-of-the-art approaches. This conclusion extensive conducted domains six benchmark databases.