Enabling self-supervised learned primal dual with Noise2Inverse

2026-06-25

AI Summary

This work addresses the challenge of supervised training in low-dose and sparse-view X-ray computed tomography (CT) reconstruction, where ground-truth images are typically unavailable. It introduces, for the first time, a self-supervised Noise2Inverse framework into the Learned Primal-Dual iterative reconstruction algorithm. By leveraging the statistical independence of measurement noise across different projection angles, the method enables effective training without requiring paired ground-truth data. The integration of Noise2Inverse’s self-supervision with the unrolled neural architecture of Learned Primal-Dual eliminates reliance on supervised labels, yielding significantly improved image quality compared to conventional reconstruction techniques and even supervised deep learning models such as U-Net, despite operating entirely without ground-truth supervision.

Abstract

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 Primal-Dual (N2I-LPD), enables training of a learned iterative reconstruction operator without ground-truth images by exploiting the statistical independence of noise in distinct measurements with respect to angular rotation of the CT-scan. We compare the proposed method with classical reconstruction methods, as well as neural network-based approaches such as a U-Net trained within the same N2I framework. The results demonstrate that N2I-LPD achieves improved reconstruction quality, highlighting the potential of combining learned reconstruction operators with self-supervised training strategies for practical CT imaging scenarios where ground-truth data is unavailable.

Problem

• X-ray computed tomography
• low-dose CT
• sparse-angle reconstruction
• self-supervised learning
• ground-truth data

Innovation

• self-supervised learning
• Learned Primal-Dual
• Noise2Inverse
• CT reconstruction
• inverse problems