At the 18th Chinese Collegiate Computing Competition (4C2025), a cross-disciplinary team from the School of Artificial Intelligence and the School of Design and Innovation at Shenzhen Technology University (SZTU) captured national attention with their project "Recoloring Relics — AI Archaeological Image Restoration System." Selected from 3,830 finalist works, the project won the national first prize, becoming one of only two gold-awarded entries from Guangdong Province in the "AI Application" track.

The award ceremony [Photo/School of Artificial Intelligence & School of Design and Innovation]

Traditional cultural relic restoration is painstaking, time-consuming, and highly dependent on the expertise of restorers. For severely damaged textiles, manuscripts, or painted artifacts, even skilled experts struggle to fully reconstruct colors and patterns lost over centuries. The SZTU team set out to address this global challenge by combining artificial intelligence with archaeology and cultural heritage studies.

The national first prize medal [Photo/School of Artificial Intelligence & School of Design and Innovation]

The Recoloring Relics system provides an end-to-end digital restoration workflow, covering image digitization, fragment stitching, damage detection, missing-area reconstruction, and visual comparison. This project goes beyond algorithmic performance, placing equal emphasis on interpretability and trustworthy evaluation within real archaeological contexts. By fostering deep collaboration between AI methods and the humanities, it advances the digital preservation and representation of cultural heritage.

Built on a curated dataset of more than 40,000 images spanning 17 dynasties and multiple artifact types, the system enables AI to learn authentic pattern logic before generating repair suggestions. Using enhanced edge-extension and feature-matching algorithms, it accurately identifies texture directions in damaged regions.

Original picture with missing areas after experts' reconstruction (left) and the restored picture with the AI system (right) [Photo/School of Artificial Intelligence & School of Design and Innovation]

The project was selected for a special exhibition during the competition due to its cultural significance and technical innovation. Experts praised its interdisciplinary approach and strong potential for real-world applications in cultural heritage protection.

From the laboratory to the national stage—and soon, to actual restoration environments—the SZTU team demonstrates how young innovators can bridge technology and the humanities to safeguard history. SZTU will continue fostering interdisciplinary collaboration to inspire more "Tech Plus Humanities" breakthroughs and cultivate versatile, forward-thinking talent for the future.

Drafted by Difien（詹进源）/ Class of 2027, College of New Materials and New Energies

Revised by Daisy（姚琦）/ International Cooperation and Exchanges Department

Edited by Brian (郑斌) / International Cooperation and Exchanges Department