Low-resolution Prior Equilibrium Network for CT Reconstruction
This work addresses incremental improvements in CT reconstruction for medical imaging, enhancing robustness in specific scenarios like sparse-view and limited-angle data.
The paper tackles the problem of unsatisfactory CT reconstruction from unrolled regularization models by introducing a low-resolution prior to improve robustness, resulting in a model that outperforms state-of-the-art methods in noise reduction, contrast-to-noise ratio, and edge preservation on sparse-view and limited-angle tasks.
The unrolling method has been investigated for learning variational models in X-ray computed tomography. However, it has been observed that directly unrolling the regularization model through gradient descent does not produce satisfactory results. In this paper, we present a novel deep learning-based CT reconstruction model, where the low-resolution image is introduced to obtain an effective regularization term for improving the network`s robustness. Our approach involves constructing the backbone network architecture by algorithm unrolling that is realized using the deep equilibrium architecture. We theoretically discuss the convergence of the proposed low-resolution prior equilibrium model and provide the conditions to guarantee convergence. Experimental results on both sparse-view and limited-angle reconstruction problems are provided, demonstrating that our end-to-end low-resolution prior equilibrium model outperforms other state-of-the-art methods in terms of noise reduction, contrast-to-noise ratio, and preservation of edge details.