Leveraging Phase Information to Boost Unrolled Network Learning for Image Deblurring

2026-06-30

AI Summary

This work addresses the common oversight in existing image deblurring methods of neglecting the critical role of phase information in fine detail recovery, which often leads to performance degradation under high noise or limited training data. To overcome this limitation, the authors propose a novel approach that explicitly models and jointly optimizes magnitude and phase components within an algorithm-unfolding network. By decomposing images into frequency-domain magnitude and phase representations, they construct an iterative optimization framework based on linear minimum mean square error (LMMSE) estimation, with statistical parameters learned end-to-end. This method transcends the conventional spatial-domain-only optimization paradigm and achieves state-of-the-art performance on benchmarks such as GoPro, RealBlur, and COCO, demonstrating particularly robust results in high-noise regimes and data-scarce scenarios.

Abstract

While most image deblurring techniques directly restore the spatial image variable, we propose an amplitude and phase decomposition recognizing the importance of accurate phase estimation in recovering sharp image details. To that end, we first develop novel linear minimum mean squared (LMMSE) estimators of the amplitude and phase of the blurred, noisy image observation. An iterative optimization algorithm follows that recovers the sharp image using the aforementioned LMMSE estimators. Finally, matrix parameters that are statistically determined and fixed in the iterative algorithm are now learned using a training dataset of clean and degraded observations. Our deblurring engine is dubbed UPADNet (Unrolled Phase and Amplitude Decomposition Network), such that each iteration of the underlying phase and amplitude recovery algorithm is parameterized and trained end-to-end. Experiments over benchmark evaluation datasets such as GoPro, RealBlur and COCO datasets confirm that UPADNet outperforms state of the art deep networks including those based on algorithm unrolling in the image domain. The benefits of UPADNet are even more pronounced in high noise and limited training data regimes.

Problem

• image deblurring
• phase estimation
• amplitude and phase decomposition
• deep learning
• algorithm unrolling

Innovation

• phase estimation
• amplitude-phase decomposition
• unrolled network
• LMMSE estimator
• image deblurring