A New Approach to Image Denoising based on Diffusion-MLP-LMMSE Scheme

In this paper, diffusion wavelet-based multiscale linear minimum mean square-error estimation (LMMSE) scheme for image denoising in conjunction to neural network is proposed, and the determination of the optimal wavelet basis with respect to the proposed scheme is also discussed. Genrally ,the over complete wavelet expansion (OWE) is more effective than the orthogonal wavelet transform (OWT) specially in image noise reduction, problems. For exploring the strong interscale dependencies of OWE, we combine the pixels at the same spatial location across scales as a vector and apply LMMSE to the vector. Compared with the LMMSE within each scale, the interscale model exploits the dependency information distributed at adjacent scales. The performance of the proposed scheme is dependent on the selection of the wavelet bases. The optimal wavelet that achieves the best tradeoffs between the two criteria can be determined from a library of wavelet bases. To estimate the wavelet coefficient statistics precisely and more accurately, we use the MLP approach of neural network which exploits the wavelet intrascale dependency and yields a local discrimination of images. The scheme improves the denoising rate as the training images are increased.