{"id":15554905,"url":"https://github.com/loseall/videosuperresolution","last_synced_at":"2025-05-15T12:04:41.475Z","repository":{"id":43901895,"uuid":"135996045","full_name":"LoSealL/VideoSuperResolution","owner":"LoSealL","description":"A collection of state-of-the-art video or single-image super-resolution architectures, reimplemented in tensorflow. ","archived":false,"fork":false,"pushed_at":"2020-09-11T14:36:45.000Z","size":9477,"stargazers_count":1657,"open_issues_count":8,"forks_count":298,"subscribers_count":53,"default_branch":"master","last_synced_at":"2025-05-15T12:04:34.864Z","etag":null,"topics":["carn","dbpn","dncnn","edsr","frvsr","ntire2019","pytorch","rcan","rdn","srcnn","srgan","srmd","super-resolution","tensorflow","vdsr","vespcn","vsr"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/LoSealL.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2018-06-04T08:42:05.000Z","updated_at":"2025-05-14T19:34:33.000Z","dependencies_parsed_at":"2022-09-17T06:50:32.736Z","dependency_job_id":null,"html_url":"https://github.com/LoSealL/VideoSuperResolution","commit_stats":null,"previous_names":[],"tags_count":17,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LoSealL%2FVideoSuperResolution","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LoSealL%2FVideoSuperResolution/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LoSealL%2FVideoSuperResolution/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LoSealL%2FVideoSuperResolution/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/LoSealL","download_url":"https://codeload.github.com/LoSealL/VideoSuperResolution/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":254337612,"owners_count":22054253,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["carn","dbpn","dncnn","edsr","frvsr","ntire2019","pytorch","rcan","rdn","srcnn","srgan","srmd","super-resolution","tensorflow","vdsr","vespcn","vsr"],"created_at":"2024-10-02T15:04:18.628Z","updated_at":"2025-05-15T12:04:36.456Z","avatar_url":"https://github.com/LoSealL.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Video Super Resolution\nA collection of state-of-the-art video or single-image super-resolution architectures, reimplemented in tensorflow.\n\n**Project uploaded to PyPI now. Try install from PyPI:**\n```shell script\npip install VSR\n```\n\n**Pretrained weights is uploading now.**\n\n**Several referenced PyTorch implementations are also included now.**\n\n**Quick Link:**\n- [Installation](#install)\n- [Getting Started](#Getting-Started)\n- [Benchmark](https://github.com/LoSealL/VideoSuperResolution/blob/master/Docs/Benchmark%20(reproduce).md)\n\n## Network list and reference (Updating)\nThe hyperlink directs to paper site, follows the official codes if the authors open sources.\n\nAll these models are implemented in **ONE** framework.\n\n|Model |Published |Code* |VSR (TF)**|VSR (Torch)|Keywords|Pretrained|\n|:-----|:---------|:-----|:---------|:----------|:-------|:---------|\n|SRCNN|[ECCV14](https://arxiv.org/abs/1501.00092)|-, [Keras](https://github.com/qobilidop/srcnn)|Y|Y| Kaiming |[√](https://github.com/LoSealL/Model/releases)|\n|RAISR|[arXiv](https://arxiv.org/abs/1606.01299)|-|-|-| Google, Pixel 3 ||\n|ESPCN|[CVPR16](https://arxiv.org/abs/1609.05158)|-, [Keras](https://github.com/qobilidop/srcnn)|Y|Y| Real time |[√](https://github.com/LoSealL/Model/releases)|\n|VDSR|[CVPR16](https://arxiv.org/abs/1511.04587)|-|Y|Y| Deep, Residual |[√](https://drive.google.com/open?id=1hW5YDxXpmjO2IfAy8f29O7yf1M3fPIg1)|\n|DRCN|[CVPR16](https://arxiv.org/abs/1511.04491)|-|Y|Y| Recurrent ||\n|DRRN|[CVPR17](http://cvlab.cse.msu.edu/pdfs/Tai_Yang_Liu_CVPR2017.pdf)|[Caffe](https://github.com/tyshiwo/DRRN_CVPR17), [PyTorch](https://github.com/jt827859032/DRRN-pytorch)|Y|Y| Recurrent ||\n|LapSRN|[CVPR17](http://vllab.ucmerced.edu/wlai24/LapSRN/)|[Matlab](https://github.com/phoenix104104/LapSRN)|Y|-| Huber loss ||\n|EDSR|[CVPR17](https://arxiv.org/abs/1707.02921)|-|Y|Y| NTIRE17 Champion |[√](https://github.com/LoSealL/Model/releases)|\n|SRGAN|[CVPR17](https://arxiv.org/abs/1609.04802)|-|Y|-| 1st proposed GAN ||\n|VESPCN|[CVPR17](https://arxiv.org/abs/1611.05250)|-|Y|Y| VideoSR |[√](https://drive.google.com/open?id=19u4YpsyThxW5dv4fhpMj7c5gZeEDKthm)|\n|MemNet|[ICCV17](https://arxiv.org/abs/1708.02209)|[Caffe](https://github.com/tyshiwo/MemNet)|Y|-|||\n|SRDenseNet|[ICCV17](http://openaccess.thecvf.com/content_ICCV_2017/papers/Tong_Image_Super-Resolution_Using_ICCV_2017_paper.pdf)|-, [PyTorch](https://github.com/wxywhu/SRDenseNet-pytorch)|Y|-| Dense |[√](https://drive.google.com/open?id=1aXAfRqZieY6mTfZUnErG84-9NfkQSeDw)|\n|SPMC|[ICCV17](https://arxiv.org/abs/1704.02738)|[Tensorflow](https://github.com/jiangsutx/SPMC_VideoSR)|T|Y| VideoSR ||\n|DnCNN|[TIP17](http://ieeexplore.ieee.org/document/7839189/)|[Matlab](https://github.com/cszn/DnCNN)|Y|Y| Denoise |[√](https://github.com/LoSealL/Model/releases)|\n|DCSCN|[arXiv](https://arxiv.org/abs/1707.05425)|[Tensorflow](https://github.com/jiny2001/dcscn-super-resolution)|Y|-|||\n|IDN|[CVPR18](https://arxiv.org/abs/1803.09454)|[Caffe](https://github.com/Zheng222/IDN-Caffe)|Y|-| Fast |[√](https://drive.google.com/open?id=1Fh3rtvrKKLAK27r518T1M_JET_LWZAFQ)|\n|RDN|[CVPR18](https://arxiv.org/abs/1802.08797)|[Torch](https://github.com/yulunzhang/RDN)|Y|-| Deep, BI-BD-DN ||\n|SRMD|[CVPR18](https://arxiv.org/abs/1712.06116)|[Matlab](https://github.com/cszn/SRMD)|-|Y| Denoise/Deblur/SR |[√](https://drive.google.com/open?id=1ORKH05-aLSbQaWB4qQulIm2INoRufuD_)|\n|DBPN|[CVPR18](https://arxiv.org/abs/1803.02735)|[PyTorch](https://github.com/alterzero/DBPN-Pytorch)|Y|Y| NTIRE18 Champion |[√](https://drive.google.com/open?id=1ymtlOjhkGmad-od0zw7yTf17nWD4KMVi)|\n|ZSSR|[CVPR18](http://www.wisdom.weizmann.ac.il/~vision/zssr/)|[Tensorflow](https://github.com/assafshocher/ZSSR)|-|-| Zero-shot ||\n|FRVSR|[CVPR18](https://arxiv.org/abs/1801.04590)|[PDF](https://github.com/msmsajjadi/FRVSR)|T|Y| VideoSR |[√](https://github.com/LoSealL/Model/releases)|\n|DUF|[CVPR18](http://openaccess.thecvf.com/content_cvpr_2018/papers/Jo_Deep_Video_Super-Resolution_CVPR_2018_paper.pdf)|[Tensorflow](https://github.com/yhjo09/VSR-DUF)|T|-| VideoSR ||\n|CARN|[ECCV18](https://arxiv.org/abs/1803.08664)|[PyTorch](https://github.com/nmhkahn/CARN-pytorch)|Y|Y| Fast |[√](https://github.com/LoSealL/Model/releases/carn)|\n|RCAN|[ECCV18](https://arxiv.org/abs/1807.02758)|[PyTorch](https://github.com/yulunzhang/RCAN)|Y|Y| Deep, BI-BD-DN ||\n|MSRN|[ECCV18](http://openaccess.thecvf.com/content_ECCV_2018/papers/Juncheng_Li_Multi-scale_Residual_Network_ECCV_2018_paper.pdf)|[PyTorch](https://github.com/MIVRC/MSRN-PyTorch)|Y|Y| |[√](https://drive.google.com/open?id=1A0LoY3oB_VnArP3GzI1ILUNJbLAEjdtJ)|\n|SRFeat|[ECCV18](http://openaccess.thecvf.com/content_ECCV_2018/papers/Seong-Jin_Park_SRFeat_Single_Image_ECCV_2018_paper.pdf)|[Tensorflow](https://github.com/HyeongseokSon1/SRFeat)|Y|Y| GAN ||\n|NLRN|[NIPS18](https://papers.nips.cc/paper/7439-non-local-recurrent-network-for-image-restoration.pdf)|[Tensorflow](https://github.com/Ding-Liu/NLRN)|T|-| Non-local, Recurrent ||\n|SRCliqueNet|[NIPS18](https://arxiv.org/abs/1809.04508)|-|-|-| Wavelet ||\n|FFDNet|[TIP18](https://ieeexplore.ieee.org/document/8365806/)|[Matlab](https://github.com/cszn/FFDNet)|Y|Y| Conditional denoise||\n|CBDNet|[CVPR19](https://arxiv.org/abs/1807.04686)|[Matlab](https://github.com/GuoShi28/CBDNet)|T|-| Blind-denoise ||\n|SOFVSR|[ACCV18](http://arxiv.org/abs/1809.08573)|[PyTorch](https://github.com/LongguangWang/SOF-VSR)|-|Y| VideoSR |[√](https://github.com/LoSealL/Model/releases/download/sofvsr/SOFVSR_x4.zip)|\n|ESRGAN|[ECCVW18](http://arxiv.org/abs/1809.00219)|[PyTorch](https://github.com/xinntao/ESRGAN)|-|Y|1st place PIRM 2018|[√](https://github.com/LoSealL/Model/releases/download/esrgan/esrgan.zip)|\n|TecoGAN|[arXiv](http://arxiv.org/abs/1811.09393)|[Tensorflow](https://github.com/thunil/TecoGAN)|-|T| VideoSR GAN|[√](https://github.com/LoSealL/Model/releases/download/tecogan/tecogan.zip)|\n|RBPN|[CVPR19](https://arxiv.org/abs/1903.10128)|[PyTorch](https://github.com/alterzero/RBPN-PyTorch)|-|Y| VideoSR |[√](https://drive.google.com/open?id=1Ozp5j-DBWJSpXY5GvxiEPKdfCaAbOXqu)|\n|DPSR|[CVPR19](https://arxiv.org/abs/1903.12529)|[Pytorch](https://github.com/cszn/DPSR)|-|-|||\n|SRFBN|[CVPR19](https://arxiv.org/abs/1903.09814)|[Pytorch](https://github.com/Paper99/SRFBN_CVPR19)|-|-||||\n|SRNTT|[CVPR19](https://arxiv.org/abs/1903.00834)|[Tensorflow](https://github.com/ZZUTK/SRNTT)|-|-|Adobe||\n|SAN|[CVPR19](http://openaccess.thecvf.com/content_CVPR_2019/papers/Dai_Second-Order_Attention_Network_for_Single_Image_Super-Resolution_CVPR_2019_paper.pdf)|[empty](https://github.com/daitao/SAN)|-|-| AliDAMO SOTA ||\n|AdaFM|[CVPR19](https://arxiv.org/abs/1904.08118)|[Pytorch](https://github.com/hejingwenhejingwen/AdaFM)|-|-| SenseTime Oral ||\n\n\\*The 1st repo is by paper author.\n\n\\**__Y__: included; __-__: not included; __T__: under-testing. \n\nYou can download pre-trained weights through [`prepare_data`](./prepare_data.py), or visit the hyperlink at **√**.\n\n## Link of datasets\n*(please contact me if any of links offend you or any one disabled)*\n\n|Name|Usage|#|Site|Comments|\n|:---|:----|:----|:---|:-----|\n|SET5|Test|5|[download](https://uofi.box.com/shared/static/kfahv87nfe8ax910l85dksyl2q212voc.zip)|[jbhuang0604](https://github.com/jbhuang0604/SelfExSR)|\n|SET14|Test|14|[download](https://uofi.box.com/shared/static/igsnfieh4lz68l926l8xbklwsnnk8we9.zip)|[jbhuang0604](https://github.com/jbhuang0604/SelfExSR)|\n|SunHay80|Test|80|[download](https://uofi.box.com/shared/static/rirohj4773jl7ef752r330rtqw23djt8.zip)|[jbhuang0604](https://github.com/jbhuang0604/SelfExSR)|\n|Urban100|Test|100|[download](https://uofi.box.com/shared/static/65upg43jjd0a4cwsiqgl6o6ixube6klm.zip)|[jbhuang0604](https://github.com/jbhuang0604/SelfExSR)|\n|VID4|Test|4|[download](https://people.csail.mit.edu/celiu/CVPR2011/videoSR.zip)|4 videos|\n|BSD100|Train|300|[download](https://uofi.box.com/shared/static/qgctsplb8txrksm9to9x01zfa4m61ngq.zip)|[jbhuang0604](https://github.com/jbhuang0604/SelfExSR)|\n|BSD300|Train/Val|300|[download](https://www2.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/segbench/BSDS300-images.tgz)|-|\n|BSD500|Train/Val|500|[download](http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/BSR/BSR_bsds500.tgz)|-|\n|91-Image|Train|91|[download](http://www.ifp.illinois.edu/~jyang29/codes/ScSR.rar)|Yang|\n|DIV2K|Train/Val|900|[website](https://data.vision.ee.ethz.ch/cvl/DIV2K/)|NTIRE17|\n|Waterloo|Train|4741|[website](https://ece.uwaterloo.ca/~k29ma/exploration/)|-|\n|MCL-V|Train|12|[website](http://mcl.usc.edu/mcl-v-database/)|12 videos|\n|GOPRO|Train/Val|33|[website](https://github.com/SeungjunNah/DeepDeblur_release)|33 videos, deblur|\n|CelebA|Train|202599|[website](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html)|Human faces|\n|Sintel|Train/Val|35|[website](http://sintel.is.tue.mpg.de/downloads)|Optical flow|\n|FlyingChairs|Train|22872|[website](https://lmb.informatik.uni-freiburg.de/resources/datasets/FlyingChairs.en.html#flyingchairs)|Optical flow|\n|DND|Test|50|[website](https://noise.visinf.tu-darmstadt.de/)|Real noisy photos|\n|RENOIR|Train|120|[website](http://ani.stat.fsu.edu/~abarbu/Renoir.html)|Real noisy photos|\n|NC|Test|60|[website](http://demo.ipol.im/demo/125/)|Noisy photos|\n|SIDD(M)|Train/Val|200|[website](https://www.eecs.yorku.ca/~kamel/sidd/)|NTIRE 2019 Real Denoise|\n|RSR|Train/Val|80|[download]()|NTIRE 2019 Real SR|\n|Vimeo-90k|Train/Test|89800|[website](http://toflow.csail.mit.edu/)|90k HQ videos|\n\nOther open datasets:\n[Kaggle](https://www.kaggle.com/datasets)\n[ImageNet](http://www.image-net.org/)\n[COCO](http://cocodataset.org/)\n\n## VSR package\nThis package offers a training and data processing framework based on [TF](https://www.tensorflow.org).\nWhat I made is a simple, easy-to-use framework without lots of encapulations and abstractions.\nMoreover, VSR can handle raw NV12/YUV as well as a sequence of images as inputs.\n\n### Install\n\n1. Prepare proper tensorflow and pytorch(optional). For example, GPU and CUDA10.0 (recommend to use `conda`):\n\n   ```shell\n   conda install tensorflow-gpu==1.15.0\n   # optional\n   # conda install pytorch\n   ```\n\n2. Install VSR package\n\n   ```bash\n   # For someone see this doc online\n   # git clone https://github.com/loseall/VideoSuperResolution \u0026\u0026 cd VideoSuperResolution\n   pip install -e .\n   ```\n### Getting Started\n\n1. Download pre-trained weights and (optinal) training datasets. For instance, let\\'s begin with VESPCN and vid4 test data:\n   ```shell\n   python prepare_data.py --filter vespcn vid4\n   ```\n   \n2. Customize backend\n   cd ~/.vsr/\n   touch config.yml\n   ```yaml\n   backend: tensorflow  # (tensorflow, pytorch)\n   verbose: info        # (debug, info, warning, error)\n   ```\n\n3. Evaluate\n   ```shell\n   cd Train\n   python eval.py srcnn -t vid4 --pretrain=/path/srcnn.pth\n   ```\n\n4. Train\n   ```shell\n   python prepare_data.py --filter mcl-v\n   cd Train\n   python train.py vespcn --dataset mcl-v --memory_limit 1GB --epochs 100\n   ```\n\nOK, that's all you need. For more details, use `--help` to get more information.\n\n----\nMore documents can be found at [Docs](https://github.com/LoSealL/VideoSuperResolution/tree/master/Docs).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Floseall%2Fvideosuperresolution","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Floseall%2Fvideosuperresolution","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Floseall%2Fvideosuperresolution/lists"}