FCL-GAN: A Lightweight and Real-Time Baseline for Unsupervised Blind Image Deblurring

TL;DR

FCL-GAN is a lightweight, real-time unsupervised deep learning model for blind image deblurring that outperforms state-of-the-art methods in speed and size without sacrificing quality.

Contribution

The paper introduces FCL-GAN, a novel unsupervised BID framework with new lightweight units, achieving significant reductions in model size and inference time while handling diverse image resolutions and domains.

Findings

FCL-GAN is 25 times lighter than SOTA models.

FCL-GAN is 5 times faster than SOTA models.

Extensive experiments show superior performance and efficiency.

Abstract

Blind image deblurring (BID) remains a challenging and significant task. Benefiting from the strong fitting ability of deep learning, paired data-driven supervised BID method has obtained great progress. However, paired data are usually synthesized by hand, and the realistic blurs are more complex than synthetic ones, which makes the supervised methods inept at modeling realistic blurs and hinders their real-world applications. As such, unsupervised deep BID method without paired data offers certain advantages, but current methods still suffer from some drawbacks, e.g., bulky model size, long inference time, and strict image resolution and domain requirements. In this paper, we propose a lightweight and real-time unsupervised BID baseline, termed Frequency-domain Contrastive Loss Constrained Lightweight CycleGAN (shortly, FCL-GAN), with attractive properties, i.e., no image domain limitation, no image resolution limitation, 25x lighter than SOTA, and 5x faster than SOTA. To guarantee the lightweight property and performance superiority, two new collaboration units called lightweight domain conversion unit(LDCU) and parameter-free frequency-domain contrastive unit(PFCU) are designed. LDCU mainly implements inter-domain conversion in lightweight manner. PFCU further explores the similarity measure, external difference and internal connection between the blurred domain and sharp domain images in frequency domain, without involving extra parameters. Extensive experiments on several image datasets demonstrate the effectiveness of our FCL-GAN in terms of performance, model size and reference time.