@article{Ye2025, 
author = {Shuainan Ye and Jianing Yin and Buwei Zhou and Tan Tang and Lingyun Yu and Ruohan Yu and Lu Jiang and Changyu Diao and Yingcai Wu},
title = {PuzzleSorter: Certainty-aware visual restoration of multiple cultural artifacts},
year = {2025},
journal = {Computational Visual Media},
volume = {11},
number = {6},
pages = {1281-1302},
keywords = {cultural heritage, fragment restoration, force-directed graph, uncertainty visualization},
url = {https://www.sciopen.com/article/10.26599/CVM.2025.9450468},
doi = {10.26599/CVM.2025.9450468},
abstract = {We present PuzzleSorter, a certainty-aware visual analytics system for cultural relic fragment restoration. Restoring cultural objects from broken fragments is a fundamental task in geometry and archaeology. Prior research proposes automatic models to classify fragments by types and assemble matched pairs successively. However, eroded fragments lead to erroneous results, posing two challenges for restorers to correct: (1) numerous fragments conceal errors within an overwhelming number of object appearances, and (2) the unknown difficulty of restoration hinders correction strategy development. To address these challenges, PuzzleSorter provides multi-criteria analysis that helps users identify certainties of current solutions and alternatives at the type, object, and fragment levels. Moreover, our system visualizes these certainties through a relation graph, which implies alternative assembly solutions with geometric context and indicates correction difficulties through neighbor proximity, number of neighbors, and path length. We demonstrate the feasibility and utility of our system through two case studies and expert interviews.}
}