Cristian Llull
A Lightweight Method for 3D Khachkar Reconstruction
Bio:
Cristián Llull is a first-year PhD student in Computer Science at the University of Chile and a part-time professor at the School of Informatics and Telecommunications in the Faculty of Engineering at Diego Portales University (UDP). He holds a Master’s degree in Computer Science and a Civil Engineering degree from the University of Chile. He has worked as a developer of on-device computer vision applications. His research focuses on 3D reconstruction methods applied to cultural heritage objects, with a special focus on the use of mobile devices.
Description of the Talk:
Recent advances in technology have significantly enhanced cultural heritage preservation, with neural networks challenging traditional 3D scanning methods for reconstruction. However, most state-of-the-art techniques remain too resource-intensive for on-device processing. In this work, we present a lightweight method for the 3D reconstruction of Khachkars; optimized for speed, accessibility, and visual fidelity. Our approach requires only a single, carefully selected image of the Khachkar to extract relief information and generate a sculpted 3D mesh closely resembling the original artifact. This technique enables realistic visualization comparable to photogrammetry, while being feasible on widely available devices such as smartphones. By lowering hardware requirements, we democratize 3D reconstruction, facilitating its use in virtual museums, educational platforms, and broader cultural dissemination.