Deep, Convergent, Unrolled Half-Quadratic Splitting for Image Deconvolution

Yanan Zhao, Yuelong Li, Haichuan Zhang, Vishal Monga, Yonina C. Eldar

Research output: Contribution to journalArticlepeer-review

Abstract

In recent years, algorithm unrolling has emerged as a powerful technique for designing interpretable neural networks based on iterative algorithms. Imaging inverse problems have particularly benefited from unrolling-based deep network design since many traditional model-based approaches rely on iterative optimization. Despite exciting progress, typical unrolling approaches heuristically design layer-specific convolution weights to improve performance. Crucially, convergence properties of the underlying iterative algorithm are lost once layer-specific parameters are learned from training data. We propose an unrolling technique that breaks the trade-off between retaining algorithm properties while simultaneously enhancing performance. We focus on image deblurring and unrolling the widely-applied Half-Quadratic Splitting (HQS) algorithm. We develop a new parametrization scheme which enforces layer-specific parameters to asymptotically approach certain fixed points. Through extensive experimental studies, we verify that our approach achieves competitive performance with state-of-the-art unrolled layer-specific learning and significantly improves over the traditional HQS algorithm. We further establish convergence of the proposed unrolled network as the number of layers approaches infinity, and characterize its convergence rate. Our experimental verification involves simulations that validate the analytical results as well as comparison with state-of-the-art non-blind deblurring techniques on benchmark datasets. The merits of the proposed convergent unrolled network are established over competing alternatives, especially in the regime of limited training.

Original languageEnglish (US)
Pages (from-to)574-588
Number of pages15
JournalIEEE Transactions on Computational Imaging
Volume10
DOIs
StatePublished - 2024

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Computer Science Applications
  • Computational Mathematics

Cite this