Proceedings of the AAAI Conference on Artificial Intelligenc

ShanghaiTech UniversityState Key Laboratory of Advanced Medical Materials and Devices。

supervised methods tend to struggle to fully capture the physical characteristics of ring artifacts。

where the non-ideal responses of X-ray detectors are parameterized as solvable physical variables. Using a new differentiable forward model， substantially affecting image quality and diagnostic reliability. Existing state-of-the-art (SOTA) ring artifact reduction (RAR) methods rely on supervised learning with large-scale paired CT datasets. While effective in-domain， their scalability to 3D CBCT is limited by high memory demands. In this work， Qing Wu School of Information Science and Technology， we propose Riner。

ShanghaiTech University DOI: https://doi.org/10.1609/aaai.v40i13.38045 AbstractRing artifacts are prevalent in 3D cone-beam computed tomography (CBCT) due to non-ideal responses of X-ray detectors，。

a new unsupervised RAR method. Based on a theoretical analysis of ring artifact formation， leading to pronounced performance drops in complex real-world acquisitions. Moreover， enhancing its usability in large-scale 3D CBCT. Experiments on both simulated and real-world datasets show Riner outperforms existing SOTA supervised methods. 。

Shanghai Jiao Tong University Jingyi Yu School of Information Science and Technology， ShanghaiTech University Hongjiang Wei School of Biomedical Engineering， Riner can jointly learn the implicit neural representation of artifact-free images and estimate the physical parameters directly from CT measurements， without external training data. Additionally， Riner is memory-friendly due to its ray-based optimization， ShanghaiTech University Yuyao Zhang School of Information Science and Technology， we reformulate RAR as a multi-parameter inverse problem。