Blind Super-Resolution With Iterative Kernel Correction

CVPR 2019 Jinjin GuHannan LuWangmeng ZuoChao Dong

Deep learning based methods have dominated super-resolution (SR) field due to their remarkable performance in terms of effectiveness and efficiency. Most of these methods assume that the blur kernel during downsampling is predefined/known (e.g., bicubic)... (read more)

PDF Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Image Super-Resolution BSD100 - 2x upscaling IKC PSNR 31.36 # 15
SSIM 0.9097 # 1
Image Super-Resolution BSD100 - 3x upscaling IKC PSNR 28.56 # 10
SSIM 0.8493 # 1
Image Super-Resolution BSD100 - 4x upscaling IKC PSNR 27.29 # 25
SSIM 0.8014 # 2
Image Super-Resolution Manga109 - 2x upscaling IKC PSNR 36.06 # 7
SSIM 0.9474 # 5
Image Super-Resolution Manga109 - 3x upscaling IKC PSNR 28.21 # 5
SSIM 0.8739 # 3
Image Super-Resolution Manga109 - 4x upscaling IKC PSNR 29.9 # 13
SSIM 0.8793 # 14
Image Super-Resolution Set14 - 2x upscaling IKC PSNR 32.82 # 14
SSIM 0.8999 # 4
Image Super-Resolution Set14 - 3x upscaling IKC PSNR 29.46 # 8
SSIM 0.8229 # 3
Image Super-Resolution Set14 - 4x upscaling IKC PSNR 28.26 # 26
SSIM 0.7688 # 33
Image Super-Resolution Set5 - 2x upscaling IKC PSNR 36.62 # 16
SSIM 0.9658 # 1
Image Super-Resolution Set5 - 3x upscaling IKC PSNR 32.16 # 12
SSIM 0.942 # 1
Image Super-Resolution Set5 - 4x upscaling IKC PSNR 31.52 # 24
SSIM 0.9278 # 2
Image Super-Resolution Urban100 - 2x upscaling IKC PSNR 30.36 # 12
SSIM 0.8949 # 4
Image Super-Resolution Urban100 - 3x upscaling IKC PSNR 25.94 # 7
SSIM 0.8165 # 2
Image Super-Resolution Urban100 - 4x upscaling IKC PSNR 25.33 # 28
SSIM 0.776 # 18

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet