From traditional craft to digital restoration: an intelligent rebirth of ancient Chinese painting restoration technique

Abstract Chinese painting restoration blends artistry and heritage preservation. While AI offers new solutions, current methods struggle to replicate its unique style and often overlook restorers' practical needs. This gap hinders the effective preservation of its historical and cultural value. To address this, our study explores AI’s potential in this field, developing InkRenew, an AI-assisted restoration system. Leveraging deep learning, InkRenew integrates dynastic and stylistic knowledge to provide real-time restoration guidance. A controlled experiment with 34 novice restorers compared AI-assisted and traditional methods, evaluating restoration quality, accuracy, and user experience. The results showed that InkRenew significantly enhanced efficiency and precision while reducing operational burden, demonstrating high usability and acceptance. Theoretically, this work bridges AI and traditional restoration, proposing a digital preservation framework. Practically, it offers restorers an efficient tool, and culturally, it advances intelligent heritage conservation. This research provides an innovative approach to safeguarding and disseminating cultural heritage.