Title: Image denoising via wavelet domain thresholding. Name: Shahriyar Matloub Augusto Roman One of the existing noise reduction techniques is based on thresholding of wavelet domain coefficients of the signal. This method was pioneered by Donoho and Johnstone in 1994. Both the soft and hard thresholding methods compare the input to a given threshold and set it to zero if its magnitude is less than the threshold. The idea is to shrink or keep (due to soft or hard thresholding) of those coefficients which are significant relative to the threshold, because they are important structures of signal, and eliminate insignificant coefficients, as they are more likely due to noise. However, we hope that for image signals it may be improved by modifying parameters of the algorithm. We suspect that two major parameters having significant impact on the performance are the threshold level and the chosen wavelet base function. In this project, we intend to implement this method, as well as explore the effects of these different parameters on performance. References [1] D. Donoho and I. Johnstone, "Ideal Spatial Adaptation via Wavelet Shrinkage," Biometrika, vol. 81, no. 3, pp. 425-455, 1994. [2] J. Liu, P.Moulin, "Complexity-Regularized Image Denoising," IEEE ICAP 1997, pp 370-373. [3] S. Mallat, A wavelet tour of signal processing. Academic Press, 1998.