A Combinatorial K-View Based algorithm for Texture Classification

Image texture classification is widely used in many applications and received considerable attention during the past decades. Several efforts have been made for developing image texture classification algorithms, including the Gray Level Co-Occurrence Matrix (GLCM), Local Binary Patterns and several K-View based algorithms. These K-View based algorithms included are K-ViewTemplate algorithm (K-View-T), K-View-Datagram algorithm (K-View-D), Fast Weighted K-View-Voting algorithm (K-View-V), K-View Using Rotation-Invariant Feature algorithm (K-View-R) and K-View Using Gray Level Co-Occurrence Matrix (K-View-G). There are some discussions about a part of these algorithms in the literatures; however, no complete experimental comparisons are made so far. In this paper, by analyzing those K-View based algorithms, an attempt to utilize the advantages of the K-View-R and K-View-V is made. The new approach which we call combinatorial K-View based method was presented. In addition, we review those K-View based algorithms and perform a comparative study based on the experiments using artificial texture images taken from the Brodatz, the evaluation method of performance between the proposed method and five different K-View algorithms are implemented by using classification accuracy, efficiency and stability.