Improved Contrastive Unpaired Translation for Metal Artifacts Reduction in Nasopharyngeal CT Images
Journal
Proceedings - 2023 IEEE Conference on Artificial Intelligence, CAI 2023
ISBN
9798350339840
Date Issued
2023-01-01
Author(s)
Abstract
Metal artifacts (MA) reduction is crucial for clinical application yet often lacks paired training data. Learning MA reduction from unpaired data and enforcing fidelity seems a trade-off. The study proposed an improved contrastive unpaired translation solution to address the issues and demonstrate its efficacy.
Subjects
constrastive learning | CT | metal artifacts reduction | negative learning
Type
conference paper