Soft Threshold Wavelet Denoising with Haar Wavelet - Implementation and Analysis

This package implements soft threshold wavelet denoising using Haar wavelet transformation. To use the program, first set the file directory as your working directory and open wavlet.fig. In the noise prompt box, input the noise intensity value between 0-0.1 (cannot be zero). Then click the process button to display experimental results, which include comparisons of the original image, noise-added image, and denoised image, along with the current PSNR value. To better understand this denoising process, let's examine each step in detail. First, set the soft threshold denoising folder as your working directory and open the wavlet.fig file. The graphical interface allows users to input noise intensity parameters through the noise prompt box, with values restricted to 0-0.1 (excluding zero) to ensure proper algorithm functioning. After clicking the process button, the system executes the denoising algorithm and displays comparative results including the original image, noise-corrupted image, and denoised output, along with quantitative PSNR measurements. The wavlet.m file contains the core program implementation, featuring key functions for Haar wavelet decomposition, soft thresholding operations, and wavelet reconstruction. The algorithm workflow includes: wavelet transformation of the input image, application of soft thresholding to wavelet coefficients, and inverse wavelet transformation to reconstruct the denoised image. By studying and running this program, users can gain deeper understanding of wavelet soft threshold denoising principles and evaluate its effectiveness through visual comparisons and PSNR metrics.