Texture Image Classification using Basic Gray Level Aura Matrices

We present a new method for texture image classification using Basic Gray Level Aura Matrices (BGLAMs). Given an unseen texture image, our approach classifies it into one of the pre-learned classes, each of which is characterized using BGLAMs. There are two stages in our algorithm: a learning stage and a classification stage. In the first stage, models of texture classes are learned from the BGLAMs of training examples using the Support Vector Machine (SVM), and in the second stage, a given texture image is classified into one of the pre-learned classes, to which the image has the largest signed distance. We compare our approach experimentally with existing approaches by performing texture classification using the Brodatz textures, the Vistex textures, and the All Sky Imager images, which are images of aurora borealis. The experimental results show that our approach has better performance than existing approaches.