Deep Learning-Guided Image Reconstruction from Incomplete Data

TL;DR

This paper introduces a novel deep learning-based iterative image reconstruction method that integrates a CNN as a quasi-projection operator, significantly improving image quality from incomplete data in limited-view scenarios.

Contribution

The paper presents a new framework combining CNNs with iterative reconstruction, inspired by proximal gradient descent, to enhance image quality from severely incomplete measurements.

Findings

Improved image quality in limited-view reconstructions.

CNN-based approach outperforms conventional methods.

Achieved state-of-the-art results in qualitative and quantitative evaluations.

Abstract

An approach to incorporate deep learning within an iterative image reconstruction framework to reconstruct images from severely incomplete measurement data is presented. Specifically, we utilize a convolutional neural network (CNN) as a quasi-projection operator within a least squares minimization procedure. The CNN is trained to encode high level information about the class of images being imaged; this information is utilized to mitigate artifacts in intermediate images produced by use of an iterative method. The structure of the method was inspired by the proximal gradient descent method, where the proximal operator is replaced by a deep CNN and the gradient descent step is generalized by use of a linear reconstruction operator. It is demonstrated that this approach improves image quality for several cases of limited-view image reconstruction and that using a CNN in an iterative method increases performance compared to conventional image reconstruction approaches. We test our method on several limited-view image reconstruction problems. Qualitative and quantitative results demonstrate state-of-the-art performance.