Image Super Resolution On Set14 4X Upscaling
Metrics
PSNR
SSIM
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | PSNR | SSIM |
|---|---|---|
| transcending-the-limit-of-local-window | 29.24 | 0.7974 |
| beyond-deep-residual-learning-for-image | 28.80 | 0.7856 |
| lightweight-and-efficient-image-super | 28.56 | 0.7803 |
| hierarchical-back-projection-network-for | 28.67 | 0.785 |
| extracter-efficient-texture-matching-with | 28.09 | 0.782 |
| hmanet-hybrid-multi-axis-aggregation-network | 29.51 | 0.8019 |
| enhancenet-single-image-super-resolution | 28.42 | 0.7774 |
| structure-preserving-super-resolution-with | 26.64 | 0.7930 |
| mair-a-locality-and-continuity-preserving | 29.2 | 0.7958 |
| a-fully-progressive-approach-to-single-image | 28.94 | - |
| feature-modulation-transformer-cross | 28.85 | 0.7872 |
| sesr-single-image-super-resolution-with | 28.32 | 0.784 |
| auto-encoded-supervision-for-perceptual-image | 27.421 | 0.7438 |
| gated-multiple-feedback-network-for-image | 28.84 | 0.7888 |
| channel-partitioned-windowed-attention-and | 29.34 | 0.7991 |
| photo-realistic-single-image-super-resolution | - | 0.7486 |
| detail-revealing-deep-video-super-resolution | 27.57 | 0.76 |
| deep-learning-based-image-super-resolution | 27.6222 | 0.7419 |
| edge-informed-single-image-super-resolution | 25.19 | 0.894 |
| progressive-perception-oriented-network-for | - | 0.7946 |
| learning-deep-cnn-denoiser-prior-for-image | 27.59 | - |
| recovering-realistic-texture-in-image-super | 26.13 | 0.694 |
| ml-craist-multi-scale-low-high-frequency | 28.4 | 0.7863 |
| deep-back-projection-networks-for-single | 29.03 | 0.791 |
| image-super-resolution-with-cross-scale-non | 28.95 | 0.7888 |
| learning-a-single-convolutional-super | 28.35 | 0.777 |
| hit-sr-hierarchical-transformer-for-efficient | 28.87 | 0.7880 |
| esrgan-enhanced-super-resolution-generative | 28.99 | 0.7917 |
| activating-more-pixels-in-image-super | 29.38 | 0.8001 |
| hit-sr-hierarchical-transformer-for-efficient | 28.84 | 0.7873 |
| channel-partitioned-windowed-attention-and | 29.36 | 0.7996 |
| cfat-unleashing-triangular-windows-for-image | 29.30 | 0.7985 |
| photo-realistic-single-image-super-resolution | 28.49 | 0.8184 |
| dual-aggregation-transformer-for-image-super | 29.29 | 0.7983 |
| activating-more-pixels-in-image-super | 29.47 | 0.8015 |
| second-order-attention-network-for-single | 29.05 | 0.7921 |
| Model 37 | 29.20 | 0.7973 |
| swinfir-revisiting-the-swinir-with-fast | 29.44 | - |
| camixersr-only-details-need-more-attention | 28.82 | 0.7870 |
| data-upcycling-knowledge-distillation-for | 28.80 | 0.7866 |
| local-texture-estimator-for-implicit | 29.06 | - |
| seven-ways-to-improve-example-based-single | 27.88 | - |
| dual-aggregation-transformer-for-image-super | 29.23 | 0.7973 |
| image-super-resolution-using-very-deep | 28.87 | 0.7889 |
| transforming-image-super-resolution-a | 28.73 | 0.7842 |
| trainable-nonlinear-reaction-diffusion-a | 27.68 | - |
| multi-scale-attention-network-for-image-super | 29.12 | 0.7941 |
| hit-sr-hierarchical-transformer-for-efficient | 28.83 | 0.7873 |
| hierarchical-information-flow-for-generalized | 29.49 | 0.8041 |
| image-super-resolution-via-dynamic-network | 28.38 | 0.7760 |
| drct-saving-image-super-resolution-away-from | 29.54 | 0.8025 |
| deep-mean-shift-priors-for-image-restoration | 26.22 | - |
| photo-realistic-single-image-super-resolution | 24.64 | 0.71 |
| fast-accurate-and-lightweight-super-1 | 28.60 | 0.7806 |
| mair-a-locality-and-continuity-preserving | 29.28 | 0.7974 |
| cascade-convolutional-neural-network-for | 28.47 | 0.7720 |
| residual-dense-network-for-image-super | 28.81 | 0.7871 |
| deeply-recursive-convolutional-network-for | 28.02 | 0.8074 |
| image-super-resolution-via-rl-csc-when | 28.29 | 0.7741 |
| spatially-adaptive-feature-modulation-for | 28.60 | 0.7813 |
| fast-and-accurate-single-image-super | 28.25 | 0.773 |
| lightweight-image-super-resolution-with-2 | 28.43 | 0.7776 |
| joint-maximum-purity-forest-with-application | 27.37 | - |
| recursive-generalization-transformer-for | 29.28 | 0.7979 |
| deep-laplacian-pyramid-networks-for-fast-and | 28.19 | 0.772 |
| image-super-resolution-via-dual-state | 28.07 | 0.770 |
| deepred-deep-image-prior-powered-by-red | 27.63 | - |
| deep-back-projection-networks-for-super | 28.82 | 0.786 |
| lightweight-image-super-resolution-with-1 | 28.58 | - |
| feedback-network-for-image-super-resolution | 28.81 | 0.7868 |
| swift-parameter-free-attention-network-for | 28.66 | 0.7834 |
| recursive-generalization-transformer-for | 29.23 | 0.7972 |
| efficient-long-range-attention-network-for | 28.96 | 0.7914 |
| single-image-super-resolution-via-a-holistic | 28.99 | 0.7907 |
| image-reconstruction-with-predictive-filter | 28.98 | 0.7904 |
| efficient-image-super-resolution-via | 28.30 | 0.7736 |
| image-super-resolution-via-attention-based | 28.94 | 0.789 |
| progressive-multi-scale-residual-network-for | 27.72 | 0.7405 |
| image-super-resolution-using-deep | 27.5 | 0.7513 |
| beyond-a-gaussian-denoiser-residual-learning | 28.04 | 0.7672 |
| enhanced-deep-residual-networks-for-single | 28.80 | 0.7876 |
| multi-scale-attention-network-for-image-super | 29.07 | 0.7934 |
| multi-level-wavelet-cnn-for-image-restoration | 28.41 | 0.7816 |
| photo-realistic-single-image-super-resolution | 25.99 | 0.7397 |
| drct-saving-image-super-resolution-away-from | 29.40 | 0.8003 |
| ram-residual-attention-module-for-single | 28.54 | 0.7800 |
| swinfir-revisiting-the-swinir-with-fast | 29.36 | 0.7993 |
| memnet-a-persistent-memory-network-for-image | 28.26 | 0.7723 |
| a-framework-for-real-time-object-detection | 28.92 | 0.7892 |
| efficient-long-range-attention-network-for | 28.78 | 0.7858 |
| real-time-single-image-and-video-super | 27.66 | 0.8004 |
| progressive-perception-oriented-network-for | 28.95 | - |
| bam-a-lightweight-and-efficient-balanced | 29.08 | 0.7925 |
| lightweight-image-super-resolution-with-2 | 28.44 | 0.7772 |
| single-image-super-resolution-with-dilated | 27.83 | 0.7631 |
| mambair-a-simple-baseline-for-image | 29.20 | 0.7961 |
| non-local-recurrent-network-for-image | 28.36 | 0.7745 |
| swinir-image-restoration-using-swin | 29.15 | 0.7958 |
| zero-shot-super-resolution-using-deep | 28.01 | 0.7651 |
| ml-craist-multi-scale-low-high-frequency | 28.53 | 0.7895 |
| densely-residual-laplacian-super-resolution | 29.02 | 0.7914 |
| image-restoration-using-convolutional-auto | 27.86 | 0.7718 |
| image-super-resolution-via-feature-augmented | 27.48 | - |