Revisiting Image Deblurring with an Efficient ConvNet

4 Feb 2023  ·  Lingyan Ruan, Mojtaba Bemana, Hans-Peter Seidel, Karol Myszkowski, Bin Chen ·

Image deblurring aims to recover the latent sharp image from its blurry counterpart and has a wide range of applications in computer vision. The Convolution Neural Networks (CNNs) have performed well in this domain for many years, and until recently an alternative network architecture, namely Transformer, has demonstrated even stronger performance. One can attribute its superiority to the multi-head self-attention (MHSA) mechanism, which offers a larger receptive field and better input content adaptability than CNNs. However, as MHSA demands high computational costs that grow quadratically with respect to the input resolution, it becomes impractical for high-resolution image deblurring tasks. In this work, we propose a unified lightweight CNN network that features a large effective receptive field (ERF) and demonstrates comparable or even better performance than Transformers while bearing less computational costs. Our key design is an efficient CNN block dubbed LaKD, equipped with a large kernel depth-wise convolution and spatial-channel mixing structure, attaining comparable or larger ERF than Transformers but with a smaller parameter scale. Specifically, we achieve +0.17dB / +0.43dB PSNR over the state-of-the-art Restormer on defocus / motion deblurring benchmark datasets with 32% fewer parameters and 39% fewer MACs. Extensive experiments demonstrate the superior performance of our network and the effectiveness of each module. Furthermore, we propose a compact and intuitive ERFMeter metric that quantitatively characterizes ERF, and shows a high correlation to the network performance. We hope this work can inspire the research community to further explore the pros and cons of CNN and Transformer architectures beyond image deblurring tasks.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Image Defocus Deblurring DPD LaKDNet Combined PSNR 26.25 # 2
Combined SSIM 0.813 # 3
LPIPS 0.153 # 1
Image Defocus Deblurring DPD (Dual-view) LaKDNet PSNR 26.89 # 4
SSIM 0.832 # 7
LPIPS 0.126 # 1
Image Deblurring GoPro LAKDNet PSNR 33.72 # 9
SSIM 0.967 # 5
Params (M) 37.5 # 14
Deblurring HIDE (trained on GOPRO) LaKDNet PSNR (sRGB) 31.58 # 3
SSIM (sRGB) 0.946 # 4
Params (M) 37.5 # 10
Deblurring RealBlur-R (trained on GoPro) LaKDNet PSNR (sRGB) 36.08 # 4
SSIM (sRGB) 0.955 # 5

Methods