Three-Level Non-Standard Wavelet Decomposition and Reconstruction

In this project, we will conduct experiments on three-level non-standard wavelet decomposition and reconstruction. We will utilize two distinct wavelet types: Haar wavelet and db9 wavelet. The implementation involves designing a program that performs 3-level wavelet decomposition and reconstruction on images. The process will be demonstrated using a personal photograph as the test subject. We will display both the decomposition and reconstruction results for each wavelet type. For Haar wavelet implementation, the program will compute reconstructions using threshold values of 0, 5, 10, and 20. The reconstructed images will be saved in PNG format, and the computational results will be recorded in an "Image Test Table". The implementation typically involves using wavelet transform functions like wavedec2 for decomposition and waverec2 for reconstruction in MATLAB, with thresholding applied to detail coefficients. For db9 wavelet processing, reconstructions will be calculated with threshold values of 0, 5, and 10. Similarly, the reconstructed images will be stored as PNG files, and results will be entered into the test table. The db9 wavelet implementation requires proper wavelet filter specification and may involve coefficient thresholding using functions like wthresh for soft or hard thresholding. These experiments will enhance understanding of wavelet decomposition/reconstruction processes and provide insights into how different threshold values affect image quality and compression performance. The code implementation typically involves multiple steps: reading the image, performing wavelet decomposition, applying thresholding to detail coefficients at each level, and reconstructing the image while evaluating metrics like PSNR or SSIM for quality assessment.