Video Super Resolution On Msu Super 1
评估指标
BSQ-rate over ERQA
BSQ-rate over LPIPS
BSQ-rate over MS-SSIM
BSQ-rate over PSNR
BSQ-rate over VMAF
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | BSQ-rate over ERQA | BSQ-rate over LPIPS | BSQ-rate over MS-SSIM | BSQ-rate over PSNR | BSQ-rate over VMAF |
|---|---|---|---|---|---|
| temporal-modulation-network-for-controllable | 1.879 | 1.377 | 0.844 | 1.481 | 1.061 |
| temporal-modulation-network-for-controllable | 13.187 | 5.015 | 4.317 | 15.144 | 3.487 |
| deep-blind-video-super-resolution | 13.145 | 13.211 | 1.438 | 6.607 | 1.383 |
| deep-blind-video-super-resolution | 15.988 | 11.435 | 0.898 | 5.765 | 0.698 |
| video-super-resolution-with-recurrent | 14.95 | 4.866 | 9.138 | 14.061 | 10.145 |
| recurrent-back-projection-network-for-video | 13.572 | 5.821 | 3.089 | 10.89 | 1.996 |
| real-esrgan-training-real-world-blind-super | 7.225 | 2.633 | 4.612 | 15.144 | 2.122 |
| real-world-super-resolution-via-kernel | 1.943 | 1.149 | 1.441 | 14.741 | 2.253 |
| 模型 9 | 16.014 | 15.363 | 4.563 | 6.473 | 6.469 |
| temporal-modulation-network-for-controllable | 21.303 | 13.988 | 1.813 | 9.43 | 1.795 |
| recurrent-back-projection-network-for-video | 1.599 | 1.335 | 0.729 | 1.127 | 0.733 |
| 模型 12 | 2.182 | 10.965 | 0.699 | 0.91 | 1.338 |
| vrt-a-video-restoration-transformer | 6.619 | 4.003 | 1.982 | 5.862 | 1.425 |
| temporal-modulation-network-for-controllable | 13.577 | 13.485 | 1.735 | 7.046 | 2.009 |
| deep-video-super-resolution-using-hr-optical | 11.458 | 4.007 | 3.566 | 8.658 | 6.596 |
| recurrent-back-projection-network-for-video | 18.314 | 11.777 | 0.884 | 5.783 | 0.689 |
| 模型 17 | 10.388 | 1.843 | 5.883 | 11.079 | 2.874 |
| local-global-fusion-network-for-video-super | 14.631 | 5.536 | 4.321 | 9.79 | 1.99 |
| temporal-modulation-network-for-controllable | 21.798 | 6.276 | 10.322 | 15.144 | 4.667 |
| local-global-fusion-network-for-video-super | 1.704 | 1.324 | 0.77 | 1.151 | 0.744 |
| video-super-resolution-with-recurrent | 13.416 | 13.232 | 5.682 | 13.403 | 6.467 |
| swinir-image-restoration-using-swin | 6.624 | 1.552 | 5.758 | 8.971 | 0.887 |
| deep-blind-video-super-resolution | 1.606 | 1.293 | 0.714 | 1.082 | 0.75 |
| deep-video-super-resolution-using-hr-optical | 15.11 | 4.034 | 7.546 | 13.076 | 7.464 |
| deep-video-super-resolution-using-hr-optical | 13.098 | 13.141 | 1.825 | 3.274 | 4.346 |
| swinir-image-restoration-using-swin | 10.854 | 4.566 | 7.105 | 15.144 | 3.32 |
| comisr-compression-informed-video-super | 11.177 | 4.801 | 11.303 | 15.144 | 10.67 |
| deep-blind-video-super-resolution | 13.476 | 4.916 | 3.886 | 10.296 | 2.093 |
| deep-blind-video-super-resolution | 7.0 | 4.371 | 2.396 | 5.845 | 1.83 |
| real-world-super-resolution-via-kernel | 1.622 | 1.206 | 1.033 | 1.064 | 1.617 |
| local-global-fusion-network-for-video-super | 18.342 | 11.759 | 0.889 | 5.768 | 1.626 |
| basicvsr-the-search-for-essential-components | 1.659 | 1.289 | 0.751 | 1.212 | 0.714 |
| 模型 33 | 11.064 | 1.219 | 1.371 | 5.718 | 0.898 |
| basicvsr-the-search-for-essential-components | 14.568 | 4.938 | 4.128 | 11.428 | 1.857 |
| real-time-super-resolution-system-of-4k-video | 12.917 | 10.748 | 5.548 | 10.701 | 6.497 |
| 模型 36 | 2.675 | 0.981 | 1.575 | 5.761 | 0.917 |
| basicvsr-the-search-for-essential-components | 8.921 | 13.198 | 1.48 | 1.906 | 1.272 |
| deep-video-super-resolution-using-hr-optical | 12.808 | 4.82 | 6.833 | 11.314 | 5.398 |
| 模型 39 | 3.679 | 1.742 | 2.861 | 6.547 | 2.308 |
| 模型 40 | 1.917 | 1.388 | 1.594 | 1.725 | 1.503 |
| recurrent-back-projection-network-for-video | 7.133 | 4.859 | 2.263 | 6.301 | 0.702 |
| recurrent-back-projection-network-for-video | 13.185 | 13.237 | 1.438 | 1.89 | 1.324 |
| 模型 43 | 21.965 | 18.057 | 5.068 | 7.372 | 7.026 |
| video-super-resolution-with-recurrent | 20.617 | 14.574 | 11.643 | 15.144 | 10.67 |
| video-super-resolution-with-recurrent | 6.58 | 10.775 | 1.023 | 13.348 | 1.5 |
| 模型 46 | 0.883 | 0.656 | 0.719 | 0.873 | 0.753 |
| real-time-super-resolution-system-of-4k-video | 10.1 | 4.0 | 8.194 | 15.144 | 10.337 |
| deep-video-super-resolution-using-hr-optical | 15.958 | 13.494 | 2.112 | 8.027 | 6.41 |
| 模型 49 | 0.922 | 0.767 | 0.643 | 0.813 | - |
| real-esrgan-training-real-world-blind-super | 5.58 | 0.733 | 0.881 | 7.874 | 0.698 |
| real-time-super-resolution-system-of-4k-video | 6.029 | 1.226 | 1.196 | 10.595 | 1.519 |
| real-world-super-resolution-via-kernel | 21.965 | 18.344 | 11.643 | 15.144 | 10.67 |
| 模型 53 | 21.965 | 17.458 | 2.888 | 8.395 | 6.797 |
| comisr-compression-informed-video-super | 13.246 | 11.026 | 6.024 | 11.497 | 8.105 |
| swinir-image-restoration-using-swin | 6.803 | 1.671 | 4.411 | 15.144 | 1.848 |
| vrt-a-video-restoration-transformer | 12.289 | 4.429 | 2.797 | 10.075 | 1.733 |
| video-super-resolution-with-recurrent | 18.327 | 13.844 | 11.643 | 15.144 | 9.796 |
| deep-video-super-resolution-using-hr-optical | 5.299 | 4.23 | 6.82 | 10.917 | 5.361 |
| deep-video-super-resolution-using-hr-optical | 18.545 | 11.236 | 4.558 | 9.07 | 3.565 |
| vrt-a-video-restoration-transformer | 18.333 | 11.496 | 0.836 | 5.777 | 0.652 |
| real-esrgan-training-real-world-blind-super | 6.328 | 12.689 | 5.393 | 8.113 | 1.464 |
| comisr-compression-informed-video-super | 3.427 | 3.851 | 7.711 | 5.761 | 9.47 |
| 模型 63 | 1.996 | 10.639 | 1.462 | 1.629 | 1.525 |
| local-global-fusion-network-for-video-super | 13.213 | 11.399 | 1.533 | 6.646 | 1.341 |
| swinir-image-restoration-using-swin | 1.575 | 1.474 | 4.641 | 8.13 | 1.304 |
| 模型 66 | 4.799 | 1.368 | 5.999 | 12.542 | 2.808 |
| real-esrgan-training-real-world-blind-super | 11.584 | 11.957 | 6.857 | 15.144 | 2.712 |
| 模型 68 | 21.965 | 16.894 | 5.735 | 12.787 | 8.635 |
| basicvsr-the-search-for-essential-components | 8.251 | 4.383 | 2.261 | 6.833 | 1.523 |
| deep-video-super-resolution-using-hr-optical | 1.544 | 1.262 | 0.843 | 2.763 | 1.213 |
| vrt-a-video-restoration-transformer | 8.92 | 11.329 | 1.257 | 6.634 | 1.217 |
| basicvsr-the-search-for-essential-components | 18.333 | 11.561 | 0.919 | 5.781 | 0.676 |
| real-world-super-resolution-via-kernel | 0.77 | 0.591 | 0.487 | 0.675 | 0.775 |
| real-esrgan-training-real-world-blind-super | 6.712 | 12.744 | 5.95 | 14.561 | 3.8 |
| deep-video-super-resolution-using-hr-optical | 4.981 | 1.26 | 0.764 | 6.058 | 1.083 |
| 模型 76 | 3.37 | 1.365 | 3.166 | 5.761 | 2.249 |
| real-time-super-resolution-system-of-4k-video | 16.733 | 5.67 | 11.643 | 15.144 | 10.67 |
| swinir-image-restoration-using-swin | 0.76 | 0.559 | 0.736 | 6.268 | 0.642 |
| local-global-fusion-network-for-video-super | 9.279 | 4.504 | 2.427 | 5.503 | 1.625 |
| deep-video-super-resolution-using-hr-optical | 18.844 | 11.273 | 4.882 | 9.245 | 4.527 |
| real-time-super-resolution-system-of-4k-video | 13.684 | 10.643 | 6.209 | 11.543 | 10.163 |
| comisr-compression-informed-video-super | 0.969 | 1.118 | 0.672 | 6.081 | 1.302 |
| vrt-a-video-restoration-transformer | 1.578 | 1.259 | 0.662 | 1.09 | 0.7 |
| real-world-super-resolution-via-kernel | 6.762 | 10.915 | 5.463 | 15.144 | 4.283 |
| comisr-compression-informed-video-super | 8.139 | 12.998 | 4.793 | 10.678 | 6.363 |