Deep Idempotent Network for Efficient Single Image Blind Deblurring

TL;DR

This paper introduces a deep idempotent network for single image blind deblurring that enhances stability and efficiency, achieving high performance with significantly reduced size and faster processing compared to existing methods.

Contribution

The work proposes a novel idempotent constraint in deblurring networks and designs a lightweight, recurrent, progressive residual network for improved and stable deblurring.

Findings

Achieves comparable high deblurring performance with much smaller model size.

Runs approximately 6.5 times faster than state-of-the-art methods.

Demonstrates superior results on synthetic and real datasets.

Abstract

Single image blind deblurring is highly ill-posed as neither the latent sharp image nor the blur kernel is known. Even though considerable progress has been made, several major difficulties remain for blind deblurring, including the trade-off between high-performance deblurring and real-time processing. Besides, we observe that current single image blind deblurring networks cannot further improve or stabilize the performance but significantly degrades the performance when re-deblurring is repeatedly applied. This implies the limitation of these networks in modeling an ideal deblurring process. In this work, we make two contributions to tackle the above difficulties: (1) We introduce the idempotent constraint into the deblurring framework and present a deep idempotent network to achieve improved blind non-uniform deblurring performance with stable re-deblurring. (2) We propose a simple yet efficient deblurring network with lightweight encoder-decoder units and a recurrent structure that can deblur images in a progressive residual fashion. Extensive experiments on synthetic and realistic datasets prove the superiority of our proposed framework. Remarkably, our proposed network is nearly 6.5X smaller and 6.4X faster than the state-of-the-art while achieving comparable high performance.