Image Super Resolution On Set14 2X Upscaling

Metrics

PSNR

SSIM

Results

Performance results of various models on this benchmark

Model Name	PSNR	SSIM	Paper Title	Repository
DRCT-L	35.36	0.9302	DRCT: Saving Image Super-resolution away from Information Bottleneck
ML-CrAIST-Li	33.64	0.9213	ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross black Attention with Image Super-resolving Transformer
CPAT	34.91	0.9277	Channel-Partitioned Windowed Attention And Frequency Learning for Single Image Super-Resolution	-
SwinOIR	33.97	0.922	Resolution Enhancement Processing on Low Quality Images Using Swin Transformer Based on Interval Dense Connection Strategy
MWCNN	33.7	-	Multi-level Wavelet-CNN for Image Restoration
MaIR	34.75	0.9268	MaIR: A Locality- and Continuity-Preserving Mamba for Image Restoration	-
CSRCNN	34.34	0.9240	Cascade Convolutional Neural Network for Image Super-Resolution	-
HAT-L	35.29	0.9293	Activating More Pixels in Image Super-Resolution Transformer
HAT	35.13	0.9282	Activating More Pixels in Image Super-Resolution Transformer
CPAT+	34.97	0.9280	Channel-Partitioned Windowed Attention And Frequency Learning for Single Image Super-Resolution	-
DRCT	34.96	0.9287	DRCT: Saving Image Super-resolution away from Information Bottleneck
DRLN+	34.43	0.9247	Densely Residual Laplacian Super-Resolution
HBPN	33.78	0.921	Hierarchical Back Projection Network for Image Super-Resolution
DRCN [[Kim et al.2016b]]	33.04	-	Deeply-Recursive Convolutional Network for Image Super-Resolution
DnCNN-3	33.03	-	Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
DBPN-RES-MR64-3	34.09	0.921	Deep Back-Projection Networks for Single Image Super-resolution
Deep CNN Denoiser	30.79	-	Learning Deep CNN Denoiser Prior for Image Restoration
HAT_FIR	35.17	-	SwinFIR: Revisiting the SwinIR with Fast Fourier Convolution and Improved Training for Image Super-Resolution
ML-CrAIST	33.77	0.922	ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross black Attention with Image Super-resolving Transformer
HAN+	34.24	0.9224	Single Image Super-Resolution via a Holistic Attention Network

0 of 35 row(s) selected.