Knowledge Base Repository
In addition to research papers, the Design Society is developing several valuable resources for those interested in the study of design. These include a repository of PhD theses, a library of case studies and transcripts of design activities, and an archive of our newsletters. Please note that these resources are accessible exclusively to Design Society members.
COG AI: A KNOWLEDGE-BASED FRAMEWORK FOR CONSTRAINED AI GENERATION IN ARTIFACTS RESTORATION
Yifei MIAO(1); Yuqi ZHAO(2); Yue BAI(3); Guanglin ZHANG(3); Jian ZHANG(1); Yuan BIAN(3)
Type:
Year:
2025
Editor:
Yong Se Kim; Yutaka Nomaguchi; Cees de Bont; Jianxi Luo; Xiaofang Yuan; Linna Hu; Meng Wang
Author:
Series:
Other endorsed
Institution:
1: 15 Tang (Tianjin) Cultural Technology Co., Ltd, China; 2: Tongji University, China; 3: Tianjin University, China
Page(s):
315-325
Abstract:
Cultural artifacts are recognized as carriers of profound historical and artistic value, and their conservation and restoration are considered essential for preserving civilizational memory and sustaining cultural heritage. In recent years, digital technologies and artificial intelligence are increasingly being utilized in the field of heritage conservation, opening new pathways for artifact preservation. However, restoration practices are often challenged by incomplete historical documentation, the irreversibility of physical interventions, and the lack of objective evaluation standards.To address these limitations, a comprehensive framework for artifact restoration is proposed. This framework is built upon a structured cultural relic database, through which an expert system is established to rigorously control the AI image generation process. The generated results are further refined through quantitative screening mechanisms. As a representative example, the implementation of the framework is illustrated through the case of porcelain artifacts, with detailed description of the entire workflow covering database construction, expert system design and feature extraction, model training, image generation, and evaluation screening.To validate the feasibility of the proposed framework, a case study was conducted using a damaged ancient book page, demonstrating the method’s effectiveness in achieving high-fidelity digital restoration.
Keywords:
