AdaRevD: Adaptive Patch Exiting Reversible Decoder Pushes the Limit of Image Deblurring

TL;DR

AdaRevD introduces an adaptive, reversible decoder architecture for image deblurring that enhances decoding capability, disentangles blur features, and speeds up processing by patch-wise classification, achieving state-of-the-art results.

Contribution

The paper proposes AdaRevD, a novel reversible decoder with adaptive patch exiting, improving deblurring performance and efficiency over existing methods.

Findings

Achieves 34.60 dB PSNR on GoPro dataset.

Disentangles high-level degradation from low-level blur.

Enables patch-wise adaptive decoding for speedup.

Abstract

Despite the recent progress in enhancing the efficacy of image deblurring, the limited decoding capability constrains the upper limit of State-Of-The-Art (SOTA) methods. This paper proposes a pioneering work, Adaptive Patch Exiting Reversible Decoder (AdaRevD), to explore their insufficient decoding capability. By inheriting the weights of the well-trained encoder, we refactor a reversible decoder which scales up the single-decoder training to multi-decoder training while remaining GPU memory-friendly. Meanwhile, we show that our reversible structure gradually disentangles high-level degradation degree and low-level blur pattern (residual of the blur image and its sharp counterpart) from compact degradation representation. Besides, due to the spatially-variant motion blur kernels, different blur patches have various deblurring difficulties. We further introduce a classifier to learn the degradation degree of image patches, enabling them to exit at different sub-decoders for speedup. Experiments show that our AdaRevD pushes the limit of image deblurring, e.g., achieving 34.60 dB in PSNR on GoPro dataset.