Image Super Resolution On Urban100 2X

PSNR

SSIM

Résultats

Résultats de performance de divers modèles sur ce benchmark

Tableau comparatif

Nom du modèle	PSNR	SSIM
a-framework-for-real-time-object-detection	32.83	0.9353
hmanet-hybrid-multi-axis-aggregation-network	35.24	0.9513
sub-pixel-back-projection-network-for	32.07	0.9277
feature-based-adaptive-contrastive	32.878	-
image-super-resolution-with-cross-scale-non	33.25	0.9386
multi-level-wavelet-cnn-for-image-restoration	32.3	-
accurate-image-super-resolution-using-very	30.76	-
activating-more-pixels-in-image-super	34.81	0.9489
drct-saving-image-super-resolution-away-from	35.17	0.9516
deep-back-projection-networks-for-single	32.92	0.935
progressive-multi-scale-residual-network-for	32.78	0.9342
densely-residual-laplacian-super-resolution	33.54	0.9402
activating-more-pixels-in-image-super	35.09	0.9505
ml-craist-multi-scale-low-high-frequency	32.93	0.9361
deeply-recursive-convolutional-network-for	30.75	-
beyond-a-gaussian-denoiser-residual-learning	30.74	-
swinfir-revisiting-the-swinir-with-fast	34.57	0.9473
hierarchical-information-flow-for-generalized	35.16	0.9505
ml-craist-multi-scale-low-high-frequency	33.04	0.937
drct-saving-image-super-resolution-away-from	34.54	0.9474
lightweight-image-super-resolution-with-1	32.17	-
channel-partitioned-windowed-attention-and	34.89	0.9487
swinfir-revisiting-the-swinir-with-fast	34.94	-
feedback-network-for-image-super-resolution	32.62	-
fast-accurate-and-lightweight-super	31.93	-
local-texture-estimator-for-implicit	33.5	-
single-image-super-resolution-via-a-holistic	33.53	0.9398
channel-partitioned-windowed-attention-and	34.76	0.9481
hierarchical-back-projection-network-for	33.12	0.938