Intriguing Findings of Frequency Selection for Image Deblurring

Authors

  • Xintian Mao Shanghai Key Laboratory of Multidimensional Information Processing, East China Normal University
  • Yiming Liu Shanghai Key Laboratory of Multidimensional Information Processing, East China Normal University
  • Fengze Liu Johns Hopkins University
  • Qingli Li Shanghai Key Laboratory of Multidimensional Information Processing, East China Normal University
  • Wei Shen MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University
  • Yan Wang Shanghai Key Laboratory of Multidimensional Information Processing, East China Normal University

DOI:

https://doi.org/10.1609/aaai.v37i2.25281

Keywords:

CV: Low Level & Physics-Based Vision, CV: Applications

Abstract

Blur was naturally analyzed in the frequency domain, by estimating the latent sharp image and the blur kernel given a blurry image. Recent progress on image deblurring always designs end-to-end architectures and aims at learning the difference between blurry and sharp image pairs from pixel-level, which inevitably overlooks the importance of blur kernels. This paper reveals an intriguing phenomenon that simply applying ReLU operation on the frequency domain of a blur image followed by inverse Fourier transform, i.e., frequency selection, provides faithful information about the blur pattern (e.g., the blur direction and blur level, implicitly shows the kernel pattern). Based on this observation, we attempt to leverage kernel-level information for image deblurring networks by inserting Fourier transform, ReLU operation, and inverse Fourier transform to the standard ResBlock. 1 × 1 convolution is further added to let the network modulate flexible thresholds for frequency selection. We term our newly built block as Res FFT-ReLU Block, which takes advantages of both kernel-level and pixel-level features via learning frequency-spatial dual-domain representations. Extensive experiments are conducted to acquire a thorough analysis on the insights of the method. Moreover, after plugging the proposed block into NAFNet, we can achieve 33.85 dB in PSNR on GoPro dataset. Our method noticeably improves backbone architectures without introducing many parameters, while maintaining low computational complexity. Code is available at https://github.com/DeepMed-Lab/DeepRFT-AAAI2023.

Downloads

Published

2023-06-26

How to Cite

Mao, X., Liu, Y., Liu, F., Li, Q., Shen, W., & Wang, Y. (2023). Intriguing Findings of Frequency Selection for Image Deblurring. Proceedings of the AAAI Conference on Artificial Intelligence, 37(2), 1905-1913. https://doi.org/10.1609/aaai.v37i2.25281

Issue

Section

AAAI Technical Track on Computer Vision II