Face Restoration of Ancient Indian Statues using Image Inpainting

Ancient sculptures of India are a significant part of our history and culture but at the same time, they get damaged because of human interference or environmental factors. Restoration of such statues, especially their faces, is highly necessary for the preservation of our rich cultural heritage. The traditional method is time-consuming and not precise. For this task, we proposed a machine learning-based technique to reconstruct missing parts of the sculpture faces from 2D Images with special focus on face reconstruction. In this work, our objective is to fill up the missing parts in facial area by using image inpainting techniques such as diffusion models. We have developed an algorithm where it detects the damaged area in the face of the statue and then applies binary masks and finally reconstructs the missing parts in the face and develops an accurate restoration of the face of ancient Indian statues. This is an easy-to-use system that allows the users to upload images of damaged statues, which are then processed by our algorithm to produce complete restored facial images. This innovative solution will help historians and archaeologists to revive our ancient Indian heritage more accurately and effectively.