High-Frequency Attention Residual GAN Network for Blind Motion Deblurring

The moving image deblurring method based on deep learning has achieved good results. However, some methods are not effective in restoring texture detail information. Therefore, this paper proposes a High-Frequency Attention Residual Module (HFAR), which is used to guide the network learn more high-frequency information improve quality of restoration. designed attention residual module consists two sub-modules, Fourier Channel (FCA) and Edge Spatial (ESA). FCA gives weight feature maps that contain multiple channels. While ESA areas details Extensive experiments different datasets show our achieves state-of-the-art performance motion deblurring.