Wavelet Domain Image Inpainting Using Nonlocal Total Variation

This paper presents the code implementation of the wavelet domain image inpainting algorithm using nonlocal total variation. The algorithm is based on the research "WAVELET INPAINTING BY NONLOCAL TOTAL VARIATION" published by Xiaoqun Zhang and Tony F. Chan in 2010 in Inverse Problems and Imaging, Volume 4, Issue 1. The implementation involves applying wavelet transform to decompose the image into different frequency components, followed by nonlocal total variation regularization to preserve texture and structure information. Key functions include wavelet decomposition/reconstruction routines and nonlocal patch-based similarity calculations. This approach effectively handles missing or damaged regions in images by leveraging both local and nonlocal image information, achieving remarkable results in the field of image restoration. The code structure typically includes wavelet coefficient thresholding, iterative optimization for the nonlocal TV minimization problem, and proper handling of boundary conditions.