Optimization of wavelet bases for texture analysis

In this paper, we propose a methodology for the optimal design of wavelet bases for maximum possible texture discrimination. The objective of this optimization process is to obtain maximum separation between local features of the texture image at different resolution scales. There are several applications which may not require reconstruction of signal from its transformed coefficients such as texture analysis, remote sensing, medical diagnostics etc. Therefore, for such applications, features are extracted at different frequency resolution scales. In case of discrimination, important information lays in the detailing coefficients at different resolution levels in contrast to lossy image compression. Therefore, our objective function is based on maximization of distinguishability function involving the computation of finer details subject to some wavelet constraints. Classification results of optimized wavelet were compared with the existing wavelet families which shows that the results obtained are superior in terms of texture distinguishability