PROJECT TITLE :

Sorted Consecutive Local Binary Pattern for Texture Classification - 2015

ABSTRACT:

In this paper, we have a tendency to propose a sorted consecutive local binary pattern (scLBP) for texture classification. Conventional strategies encode only patterns whose spatial transitions are no more than two, whereas scLBP encodes patterns no matter their spatial transition. Standard ways don't encode patterns on account of rotation-invariant encoding; on the opposite hand, patterns with a lot of than two spatial transitions have discriminative power. The proposed scLBP encodes all patterns with any range of spatial transitions whereas maintaining their rotation-invariant nature by sorting the consecutive patterns. Furthermore, we have a tendency to introduce dictionary learning of scLBP based mostly on kd-tree that separates data with a space partitioning strategy. Since the elements of sorted consecutive patterns lie in different space, it can be generated to a discriminative code with kd-tree. Finally, we have a tendency to gift a framework in which scLBPs and the kd-tree can be combined and utilised. The results of experimental analysis on 5 texture information sets-Outex, CUReT, UIUC, UMD, and KTH-TIPS2-a-indicate that our proposed framework achieves the most effective classification rate on the CUReT, UMD, and KTH-TIPS2-a knowledge sets compared with conventional ways. The results additionally indicate that only a marginal distinction exists between the most effective classification rate of conventional ways which of the proposed framework on the UIUC and Outex data sets.