Ancient Mural Image Region Regeneration and Filling Using Patch Matching Based on NonExistent Elements Priority
Sujata G. Bhele, Shashank Shriramwar, Poonam T. Agarkar · 2023
Ancient Mural Images (AMI) are one of the most important treasures that signify the culture and heritage of any region. AMI also reflects the information regarding the artistic values, art techniques, and materials along with evidence of rich history and culture. Every small or big nation had felt the thirst to conserve their ancestor's wealth carried from decades of sustaining natural and artificial attacks. As a part of reconstructing and repair, the work presented in this paper presents a Patch-Matching Algorithm (PMA) based on Non-Existent Elements Priority (NEEP) to regenerate the missing regions and fill the background using the known information in the mural image. Two statistical approaches for regeneration and filling are used respectively to make the construction resemble the degraded image. The NEEP mechanism considers the patch to be constructed or filled first and relies on several elements to recover. The color texture-structure conserving PMA reproduces the degraded regions in the mural images with a high peak signal-to-noise ratio (PSNR). The inpainting, region filling, and reconstruction quality of the proposed PMA-NEEP scheme showed superior performance in terms of perception when tested over low-resolution texture-structure images at the cost of computational complexity.