An Neural Based Approach for Optimizing Diffusion Wavlets Cofficients (An Application to image Denoising)

Abstract: In this paper ,a special class of wavelets known as Diffusion Wavelets are optimized through their coefficients, the desired process is started by applying the diffusion process to a noisy signal in our case the noisy image. We apply the neural network MLP for the determination of the optimal wavelet basis. Further we fetch the benefit of OWE (over complete wavelet expansion ) over OWT(orthogonal wavelet transform) specially in image diminution problems. The performance of the proposed method is dependent on selection of the wavelet basis.The desired scheme yield fruitful results in image denoising