Digital Restoration of Ancient Murals: Assessing the Efficacy of Coherent Transport Inpainting with Damage Ratio Analysis

Preserving and restoring ancient murals is a critical task to maintain the cultural heritage. However, the ancient murals often suffer from degradation over time, including fading, discoloration, and damages. Image inpainting is the process of repairing the damage or missing areas in an image to restore the visual appearance. In this paper, we present an efficient image restoration framework for ancient Myanmar murals that analyze the features of mural paintings and enhance the murals, automatically extract damage regions, inpaint with coherent transport method and finally, modify the inpainted result with post-processing. The nature of the murals is different; therefore, the proposed technique initiates with a comprehensive analysis on the input mural image and to maintain the global and local consistency of the whole image, the framework is finalized with the post-processing. The evaluation of the inpainting process is done with the calculation of damage ratio analysis that is designed for non-reference image quality assessment. Experiments are conducted with the murals from Po Win Taung and Kyauk Taw Gyi Pagoda to analyze and derive meaningful insights. The experimental result shows that the computational efficiency and scalability of the proposed framework make it suitable for ancient Myanmar murals restoration.