Image Texture Classification Using Texture Spectrum and Local Binary Pattern

Abstract

Spatial image classifier which incorporates contextual information to classify each pixel in the raw images has been used widely in texture analysis. The spatial classifier strives to capture the spatial relationships encoded in aerial photographs, textural and natural images. In this paper, we aimed to analyze and compare some of the simple but powerful spatial image classifiers to explore their strengths and weaknesses in remote sensing applications. Specifically, Texture Spectrum (TS) and Local Binary Pattern (LBP) will be compared. These features are easy to compute and yet useful in discriminating different patterns of textures. Co-occurrence probabilities (GLCPs) are used as the benchmark for the evaluation. There are some reviews and discussions about these methods in the literature; however, no experimental comparisons are made so far. Experimental results will be provided in this report.