Learning Single Image Defocus Deblurring with Misaligned Training Pairs

TL;DR

This paper introduces a joint deblurring and reblurring framework that effectively handles misaligned training pairs in single image defocus deblurring, improving both accuracy and visual quality.

Contribution

The proposed JDRL framework incorporates a bi-directional optical flow and reblurring modules to address misalignment issues in training data for defocus deblurring.

Findings

Improves deblurring accuracy on multiple datasets

Reduces deformation artifacts caused by misaligned training pairs

Establishes a new dataset for defocus deblurring research

Abstract

By adopting popular pixel-wise loss, existing methods for defocus deblurring heavily rely on well aligned training image pairs. Although training pairs of ground-truth and blurry images are carefully collected, e.g., DPDD dataset, misalignment is inevitable between training pairs, making existing methods possibly suffer from deformation artifacts. In this paper, we propose a joint deblurring and reblurring learning (JDRL) framework for single image defocus deblurring with misaligned training pairs. Generally, JDRL consists of a deblurring module and a spatially invariant reblurring module, by which deblurred result can be adaptively supervised by ground-truth image to recover sharp textures while maintaining spatial consistency with the blurry image. First, in the deblurring module, a bi-directional optical flow-based deformation is introduced to tolerate spatial misalignment between deblurred and ground-truth images. Second, in the reblurring module, deblurred result is reblurred to be spatially aligned with blurry image, by predicting a set of isotropic blur kernels and weighting maps. Moreover, we establish a new single image defocus deblurring (SDD) dataset, further validating our JDRL and also benefiting future research. Our JDRL can be applied to boost defocus deblurring networks in terms of both quantitative metrics and visual quality on DPDD, RealDOF and our SDD datasets.