Urban100

Introduced by Huang et al. in Single Image Super-Resolution From Transformed Self-Exemplars

The Urban100 dataset contains 100 images of urban scenes. It commonly used as a test set to evaluate the performance of super-resolution models.

Homepage

Task	Dataset Variant	Best Model
Image Super-Resolution	Urban100 - 4x upscaling	DRCT-L
Image Super-Resolution	Urban100 - 2x upscaling	HAT-L
Image Super-Resolution	Urban100 - 3x upscaling	HAT-L
Grayscale Image Denoising	Urban100 sigma50	Restormer-Local
Grayscale Image Denoising	Urban100 sigma25	Restormer-Local
Color Image Denoising	Urban100 sigma50	Restormer-Local
Color Image Denoising	urban100 sigma15	KBNet
Image Super-Resolution	Urban100 - 8x upscaling	DRLN+
Blind Super-Resolution	Urban100 - 2x upscaling	DCLS
Color Image Denoising	Urban100 sigma25	KBNet
Blind Super-Resolution	Urban100 - 4x upscaling	DCLS
Grayscale Image Denoising	urban100 sigma15	Restormer-Local
Grayscale Image Denoising	Urban100 sigma70	Residual Dense Network +
Color Image Denoising	Urban100 sigma10	Residual Dense Network +
Color Image Denoising	Urban100 sigma30	Restormer-Local
Blind Super-Resolution	Urban100 - 3x upscaling	DCLS
Grayscale Image Denoising	Urban100 sigma10	Residual Dense Network +
Color Image Denoising	Urban100 sigma70	Residual Dense Network +
Image Denoising	urban100 sigma15	SCUNet SCUNet
Image Denoising	Urban100 sigma50	SCUNet SCUNet
Image Super-Resolution	Urban100 - 16x upscaling	ABPN
Compressive Sensing	Urban100 - 2x upscaling	AMPA-Net
Joint Demosaicing and Denoising	Urban100	PIPNet
Grayscale Image Denoising	Urban100 sigma30	Residual Dense Network +