Ring-artifacts removal for spectral photon counting CT

Compared to energy-integrating detectors, photon counting detectors (PCDs) offer a better spatial resolution, higher contrast-to-noise ratio, elimination of electronic noise, improved dose efficiency, and routine multi-energy imaging. However, limited by current processing technologies, the removal of ring artifacts has become an inevitable research challenge in the pursuit of high-quality imaging. In this paper, we propose a dual-domain optimization model that integrates dual-spectral imaging with ring artifact removal. The model incorporates unidirectional total variation (TV) regularization, group sparsity regularization, and TV regularization to enable artifact removal during basis material decomposition. The first two relate the inconsistency of the detector response to the scanning angle, which can correct the detector response more accurately. The latter exploits the spatial piecewise smooth structure of the clean image to further remove the residual ring artifacts. In addition, an efficient alternating minimization scheme is designed to solve the proposed model. Extensive experiments on both simulated and real data demonstrate that our method achieves an optimal balance between artifact suppression and detail preservation. Additionally, it outperforms several state-of-the-art ring artifact removal methods in both quantitative and qualitative evaluations.