An efficient wavelet based approach for texture classification with feature analysis

Textures play important roles in many image processing applications, since images of real objects often do not exhibit regions of uniform and smooth intensities, but variations of intensities with certain repeated structures or patterns, referred to as visual texture. The textural patterns or structures mainly result from the physical surface properties, such as roughness or oriented structured of a tactile quality. It is widely recognized that a visual texture, which can easily perceive, is very difficult to define. The difficulty results mainly from the fact that different people can define textures in applications dependent ways or with different perceptual motivations, and they are not generally agreed upon single definition of texture [1]. The development in multi-resolution analysis such as Gabor and wavelet transform help to overcome this difficulty [2]. In this paper it describes that, texture classification using Wavelet Statistical Features (WSF), Wavelet Co-occurrence Features (WCF) and to combine both the features namely Wavelet Statistical Features and Wavelet Co-occurrence Features of wavelet transformed images with different feature databases can results better [2]. And further the Features are analyzed introducing Noise (Gaussian, Poisson, Salt n Paper and Speckle) in the image to be classified. The result suggests that the efficiency of Wavelet Statistical Feature is higher in classification even in noise as compared to other Features efficiency. Wavelet based decomposing is used to classify the image with code prepared in MATLAB.