Deep Constrained Least Squares for Blind Image Super-Resolution

In this paper, we tackle the problem of blind image super-resolution(SR) with a reformulated degradation model and two novel modules. Following the common practices of blind SR, our method proposes to improve both the kernel estimation as well as the kernel-based high-resolution image restoration. To be more specific, we first reformulate the degradation model such that the deblurring kernel estimation can be transferred into the low-resolution space. On top of this, we introduce a dynamic deep linear filter module. Instead of learning a fixed kernel for all images, it can adaptively generate deblurring kernel weights conditional on the input and yield a more robust kernel estimation. Subsequently, a deep constrained least square filtering module is applied to generate clean features based on the reformulation and estimated kernel. The deblurred feature and the low input image feature are then fed into a dual-path structured SR network and restore the final high-resolution result. To evaluate our method, we further conduct evaluations on several benchmarks, including Gaussian8 and DIV2KRK. Our experiments demonstrate that the proposed method achieves better accuracy and visual improvements against state-of-the-art methods.

PDF Abstract CVPR 2022 PDF CVPR 2022 Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Blind Super-Resolution BSD100 - 2x upscaling DCLS PSNR 32.04 # 1
SSIM 0.8907 # 1
Blind Super-Resolution BSD100 - 3x upscaling DCLS PSNR 29.07 # 1
SSIM 0.7956 # 2
Blind Super-Resolution BSD100 - 4x upscaling DCLS PSNR 27.6 # 1
SSIM 0.7285 # 2
Blind Super-Resolution DIV2KRK - 2x upscaling DCLS PSNR 32.75 # 1
SSIM 0.9094 # 1
Blind Super-Resolution DIV2KRK - 4x upscaling DCLS PSNR 28.99 # 1
SSIM 0.7946 # 1
Blind Super-Resolution Manga109 - 2x upscaling DCLS PSNR 38.31 # 1
SSIM 0.974 # 1
Blind Super-Resolution Manga109 - 3x upscaling DCLS PSNR 33.54 # 1
SSIM 0.9414 # 1
Blind Super-Resolution Manga109 - 4x upscaling DCLS PSNR 30.86 # 1
SSIM 0.9086 # 1
Blind Super-Resolution Set14 - 2x upscaling DCLS PSNR 33.46 # 1
SSIM 0.9103 # 1
Blind Super-Resolution Set14 - 3x upscaling DCLS PSNR 30.29 # 1
SSIM 0.8329 # 1
Blind Super-Resolution Set14 - 4x upscaling DCLS PSNR 28.54 # 1
SSIM 0.7728 # 1
Blind Super-Resolution Set5 - 2x upscaling DCLS PSNR 37.63 # 1
SSIM 0.9554 # 2
Blind Super-Resolution Set5 - 3x upscaling DCLS PSNR 34.21 # 1
SSIM 0.9218 # 2
Blind Super-Resolution Set5 - 4x upscaling DCLS PSNR 32.12 # 1
SSIM 0.889 # 2
Blind Super-Resolution Urban100 - 2x upscaling DCLS PSNR 31.69 # 1
SSIM 0.9202 # 1
Blind Super-Resolution Urban100 - 3x upscaling DCLS PSNR 28.03 # 1
SSIM 0.8444 # 1
Blind Super-Resolution Urban100 - 4x upscaling DCLS PSNR 26.15 # 1
SSIM 0.7809 # 1

Methods


No methods listed for this paper. Add relevant methods here