Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Blur2Blur: Blur Conversion for Unsupervised Image Deblurring on Unknown Domains

Bang-Dang Pham, Phong Tran, Anh Tran, Cuong Pham, Rang Nguyen, Minh Hoai

TL;DR

Blur2Blur tackles camera-specific image deblurring by learning a blur-to-blur translator that maps unknown blur to a known blur domain with a pretrained deburring model. This two-stage approach reduces the learning burden by focusing on blur patterns rather than reconstructing fine details, enabling effective deblurring with unpaired data. The method combines adversarial and perceptual losses across multiple scales and relies on a carefully chosen Known Blur domain and blur-kernel transfer to synthesize training data. Across real-world and synthetic benchmarks, Blur2Blur delivers substantial PSNR gains and demonstrates practical applicability in real devices and video scenarios.

Abstract

This paper presents an innovative framework designed to train an image deblurring algorithm tailored to a specific camera device. This algorithm works by transforming a blurry input image, which is challenging to deblur, into another blurry image that is more amenable to deblurring. The transformation process, from one blurry state to another, leverages unpaired data consisting of sharp and blurry images captured by the target camera device. Learning this blur-to-blur transformation is inherently simpler than direct blur-to-sharp conversion, as it primarily involves modifying blur patterns rather than the intricate task of reconstructing fine image details. The efficacy of the proposed approach has been demonstrated through comprehensive experiments on various benchmarks, where it significantly outperforms state-of-the-art methods both quantitatively and qualitatively. Our code and data are available at https://zero1778.github.io/blur2blur/

Blur2Blur: Blur Conversion for Unsupervised Image Deblurring on Unknown Domains

TL;DR

Blur2Blur tackles camera-specific image deblurring by learning a blur-to-blur translator that maps unknown blur to a known blur domain with a pretrained deburring model. This two-stage approach reduces the learning burden by focusing on blur patterns rather than reconstructing fine details, enabling effective deblurring with unpaired data. The method combines adversarial and perceptual losses across multiple scales and relies on a carefully chosen Known Blur domain and blur-kernel transfer to synthesize training data. Across real-world and synthetic benchmarks, Blur2Blur delivers substantial PSNR gains and demonstrates practical applicability in real devices and video scenarios.

Abstract

This paper presents an innovative framework designed to train an image deblurring algorithm tailored to a specific camera device. This algorithm works by transforming a blurry input image, which is challenging to deblur, into another blurry image that is more amenable to deblurring. The transformation process, from one blurry state to another, leverages unpaired data consisting of sharp and blurry images captured by the target camera device. Learning this blur-to-blur transformation is inherently simpler than direct blur-to-sharp conversion, as it primarily involves modifying blur patterns rather than the intricate task of reconstructing fine image details. The efficacy of the proposed approach has been demonstrated through comprehensive experiments on various benchmarks, where it significantly outperforms state-of-the-art methods both quantitatively and qualitatively. Our code and data are available at https://zero1778.github.io/blur2blur/
Paper Structure (25 sections, 7 equations, 11 figures, 6 tables)

This paper contains 25 sections, 7 equations, 11 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Approach Overview
  5. Blur-to-blur translation
  6. Known Blur Selection
  7. Experiments
  8. Experimental Setups
  9. Datasets and implementation details
  10. Baselines
  11. Image Deblurring Results
  12. Blur2Blur Visualization
  13. Ablation Study for Blur to Sharp Ratio
  14. Practicality Evaluation
  15. Conclusions
  16. ...and 10 more sections

Figures (11)

  • Figure 1: We address the unsupervised image deblurring problem by training a blur translator that converts an input image with unknown blur to an image with a predefined known blur. The figure shows the effectiveness of our approach. The blurry images before and after translation (left image in each box) exhibit similar visual content but have different blur patterns (zoomed-in patches). While a standard image deblurring technique fails to restore the unknown-blur image, it successfully recovers the known-blur version, yielding an approximate 2.2 dB increase in PSNR score (noted below each deblurred image on the right side of each box).
  • Figure 2: Unknown- and known-blur datasets
  • Figure 3: Blur translation
  • Figure 5: Comparing image deblurring results on three benchmark datasets with NAFNet. Due to space limit, we skip the results with Restormer backbone, which is similar but slightly worse than those with NAFNet. Best viewed when magnified on a digital display.
  • Figure 6: (a) A comparison of original images and their corresponding Blur2Blur converted version; (b) Selected examples demonstrating the GoPro dataset's blur pattern (Zoom for best view).
  • ...and 6 more figures