Adaptive Single Image Deblurring

TL;DR

This paper introduces an efficient, adaptive deep learning method for dynamic scene image deblurring that effectively handles large blur variations by combining pixel-wise and feature attention mechanisms with global-local filtering.

Contribution

It proposes a novel content-aware global-local filtering module and a patch hierarchical attentive architecture for improved non-uniform motion deblurring.

Findings

Outperforms prior methods on deblurring benchmarks

Achieves better quality with fewer parameters

Demonstrates robustness to large blur variations

Abstract

This paper tackles the problem of dynamic scene deblurring. Although end-to-end fully convolutional designs have recently advanced the state-of-the-art in non-uniform motion deblurring, their performance-complexity trade-off is still sub-optimal. Existing approaches achieve a large receptive field by a simple increment in the number of generic convolution layers, kernel-size, which comes with the burden of the increase in model size and inference speed. In this work, we propose an efficient pixel adaptive and feature attentive design for handling large blur variations within and across different images. We also propose an effective content-aware global-local filtering module that significantly improves the performance by considering not only the global dependencies of the pixel but also dynamically using the neighboring pixels. We use a patch hierarchical attentive architecture composed of the above module that implicitly discover the spatial variations in the blur present in the input image and in turn perform local and global modulation of intermediate features. Extensive qualitative and quantitative comparisons with prior art on deblurring benchmarks demonstrate the superiority of the proposed network.