Deep unfolding method improves image reconstruction accuracy

Newton Deep Unfolding for Compressed Sensing

Computer Vision and Pattern Recognition

Summary

Reconstructing images from limited data is hard and often slow or inaccurate. The authors propose a new technique that uses a more advanced form of math, called second-order optimization, to guide the reconstruction process better. Their method creates intermediate steps that improve the image quality as it recovers. Tests show their approach works well even when only a few measurements are available and adapts better to the reconstruction progress.

What this means in practice

Authors

Changhua He, Xianchao Xiu

Abstract

Compressed sensing (CS) reconstructs images from highly limited measurements, but existing deep unfolding methods are typically driven by first-order optimization and weakly exploit the optimization states generated during reconstruction. To address these limitations, we propose a Newton deep unfolding network (NDU-Net), which, to the best of our knowledge, is the first deep unfolding framework that leverages second-order optimization for CS reconstruction. Specifically, NDU-Net introduces a Newton update (NU) module to estimate Newton-type update directions and generate optimization states that characterize the current reconstruction process. Furthermore, a Newton-guided multi-scale prior (MP) module is designed to incorporate these optimization states into multi-scale feature restoration, thereby enabling the learned prior to adapt to the current reconstruction stage. Experimental results under different CS ratios confirm that our proposed NDU-Net achieves promising reconstruction performance and exhibits enhanced robustness. Our code is available at https://github.com/xianchaoxiu/DNU-Net.