{"id":"arxiv/2606.26991","url":"https://arxiv.org/abs/2606.26991","tags":["eess.IV","cs.LG","math.OC"],"score":0,"title":"Enabling self-supervised learned primal dual with Noise2Inverse","domain":"eess.IV","source":"arxiv","authors":["Antti Sällinen","Siiri Rautio","Santeri Kaupinmäki","Andreas Hauptmann"],"reviews":0,"subject":"eess.IV","summary":"X-ray computed tomography reconstruction is an ill-posed inverse problem, particularly in low-dose and sparse-angle settings where measurements are noisy and incomplete. While learned reconstruction methods such as the Learned Primal-Dual algorithm achieve strong performance, they typically rely on supervised training with access to ground-truth data, which is often unavailable in practice. In this work, we propose a self-supervised reconstruction method by extending the Noise2Inverse framework to the Learned Primal-Dual algorithm. The resulting approach, called Noise2Inverse Learned…","category":"research","comments":0,"disputed":false,"division":"eess","mentions":0,"discussed":0,"published":"2026-06-25","reactions":{}}