Image denoising with optimized subband threshold

Denoising is basic work for image processing. In this paper, a method based on subband threshold for image denoising is described. Thresholding is a nonlinear technique and it is very simple. But how to estimate threshold is a key problem. After studying the relationship between threshold and PSNR of image, an algorithm for estimating an optimal threshold is proposed. Results of simulation show that the best performance in image denoising may be achieved by the "optimization threshold" technique.