Fusion of optimized moment based and Gabor texture features for better texture classification

Feature extraction is one of the most vital steps involved in image description. Every feature extraction technique has its own merits and demerits. For a particular application a carefully worked fusion of features, extracted using different techniques, can enhance their image description capabilities. Optimized moment features, obtained from author's previous work, have shown promising results in classifying the textures having reasonable variance in periodicity of patterns and identical second order statistics. Gabor feature extraction is an established technique in describing the texture having features in the range of low frequencies. However, in the presence of periodic variance or impulsive noise, the Gabor filters generate highly variable features at higher frequency. This paper explores the fusion of optimized moment and Gabor energy texture features. The Fisher linear discriminant analysis shows that the discrimination effectiveness of the features increases after fusion. Results have also been validated experimentally through the classification of real texture images.