Image Super Resolution On Urban100 4X
평가 지표
PSNR
SSIM
평가 결과
이 벤치마크에서 각 모델의 성능 결과
비교 표
| 모델 이름 | PSNR | SSIM |
|---|---|---|
| lightweight-and-efficient-image-super | 26.03 | 0.7835 |
| image-super-resolution-by-neural-texture | 25.5 | - |
| non-local-recurrent-network-for-image | 25.79 | 0.7729 |
| deep-back-projection-networks-for-single | 27.08 | 0.814 |
| swinfir-revisiting-the-swinir-with-fast | 28.43 | - |
| fast-and-accurate-single-image-super | 25.41 | 0.7632 |
| deep-laplacian-pyramid-networks-for-fast-and | 25.21 | 0.756 |
| single-image-super-resolution-via-a-holistic | 27.02 | 0.8131 |
| hierarchical-back-projection-network-for | 27.3 | 0.818 |
| swinir-image-restoration-using-swin | 27.45 | 0.8254 |
| image-super-resolution-with-cross-scale-non | 27.22 | 0.8168 |
| esrgan-enhanced-super-resolution-generative | 27.03 | 0.8153 |
| data-upcycling-knowledge-distillation-for | 26.43 | 0.7972 |
| multi-step-reinforcement-learning-for-single | 23.28 | 0.7517 |
| a-framework-for-real-time-object-detection | 26.74 | 0.806 |
| hmanet-hybrid-multi-axis-aggregation-network | 28.69 | 0.8512 |
| residual-dense-network-for-image-super | 26.61 | 0.8028 |
| multi-level-wavelet-cnn-for-image-restoration | 26.27 | 0.7890 |
| mair-a-locality-and-continuity-preserving | 27.89 | 0.8336 |
| enhancenet-single-image-super-resolution | 25.66 | 0.7703 |
| drct-saving-image-super-resolution-away-from | 28.70 | 0.8508 |
| memnet-a-persistent-memory-network-for-image | 25.50 | 0.7630 |
| image-super-resolution-via-rl-csc-when | 25.59 | 0.7680 |
| perception-oriented-single-image-super | 24.33 | 0.7707 |
| activating-more-pixels-in-image-super | 28.60 | 0.8498 |
| structure-preserving-super-resolution-with | 24.799 | 0.9481 |
| second-order-attention-network-for-single | 27.23 | 0.8169 |
| beyond-deep-residual-learning-for-image | 26.42 | 0.7940 |
| ram-residual-attention-module-for-single | 26.05 | 0.7834 |
| hierarchical-information-flow-for-generalized | 28.72 | 0.8514 |
| mair-a-locality-and-continuity-preserving | 27.71 | 0.8305 |
| deep-back-projection-networks-for-super | - | 0.795 |
| image-super-resolution-via-dual-state | 25.08 | 0.747 |
| data-upcycling-knowledge-distillation-for | 26.62 | 0.802 |
| image-super-resolution-using-very-deep | 26.82 | 0.8087 |
| lightweight-image-super-resolution-with-1 | 26.04 | - |
| channel-partitioned-windowed-attention-and | 28.22 | 0.8408 |
| ml-craist-multi-scale-low-high-frequency | 26.68 | 0.8057 |
| a-fully-progressive-approach-to-single-image | 26.89 | - |
| feedback-network-for-image-super-resolution | 26.6 | 0.8015 |
| channel-partitioned-windowed-attention-and | 28.33 | 0.8425 |
| image-super-resolution-using-deep | 24.52 | 0.7221 |
| data-upcycling-knowledge-distillation-for | 26.45 | 0.7963 |
| progressive-perception-oriented-network-for | - | 0.8169 |
| image-processing-gnn-breaking-rigidity-in | 28.13 | 0.8392 |
| fast-accurate-and-lightweight-super-1 | 26.07 | 0.7837 |
| densely-residual-laplacian-super-resolution | 27.14 | 0.8149 |
| feature-based-adaptive-contrastive | 26.606 | - |
| sesr-single-image-super-resolution-with | 25.42 | 0.771 |
| auto-encoded-supervision-for-perceptual-image | 26.148 | 0.7884 |
| esrgan-enhanced-super-resolution-generative | 23.14 | 0.6577 |
| one-to-many-approach-for-improving-super | - | - |
| swinfir-revisiting-the-swinir-with-fast | 28.12 | 0.8393 |
| progressive-perception-oriented-network-for | 27.01 | - |
| drct-saving-image-super-resolution-away-from | 28.40 | 0.8457 |
| gated-multiple-feedback-network-for-image | 26.69 | 0.8048 |
| beyond-a-gaussian-denoiser-residual-learning | 25.2 | 0.7521 |
| learning-a-single-convolutional-super | 25.68 | 0.773 |
| image-super-resolution-via-attention-based | 27.06 | 0.811 |
| ml-craist-multi-scale-low-high-frequency | 26.53 | 0.8019 |
| enhanced-deep-residual-networks-for-single | 26.64 | 0.8033 |
| extracter-efficient-texture-matching-with | 26.04 | 0.785 |
| activating-more-pixels-in-image-super | 28.37 | 0.8447 |
| local-texture-estimator-for-implicit | 27.24 | - |