{"id":32986025,"url":"https://github.com/SYSU-SAIL/PASSRnet","last_synced_at":"2025-11-18T02:02:06.195Z","repository":{"id":105884363,"uuid":"172455888","full_name":"SYSU-SAIL/PASSRnet","owner":"SYSU-SAIL","description":"[CVPR 2019] Learning Parallax Attention for Stereo Image Super-Resolution","archived":false,"fork":false,"pushed_at":"2020-12-04T13:39:39.000Z","size":15001,"stargazers_count":303,"open_issues_count":7,"forks_count":67,"subscribers_count":9,"default_branch":"master","last_synced_at":"2025-05-24T06:12:21.448Z","etag":null,"topics":["depth-estimation","parallax","stereo-matching","super-resolution"],"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/SYSU-SAIL.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2019-02-25T07:27:53.000Z","updated_at":"2025-03-06T01:47:43.000Z","dependencies_parsed_at":"2023-04-13T04:21:57.318Z","dependency_job_id":null,"html_url":"https://github.com/SYSU-SAIL/PASSRnet","commit_stats":null,"previous_names":["sysu-sail/passrnet","the-learning-and-vision-atelier-lava/passrnet"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/SYSU-SAIL/PASSRnet","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SYSU-SAIL%2FPASSRnet","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SYSU-SAIL%2FPASSRnet/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SYSU-SAIL%2FPASSRnet/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SYSU-SAIL%2FPASSRnet/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/SYSU-SAIL","download_url":"https://codeload.github.com/SYSU-SAIL/PASSRnet/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SYSU-SAIL%2FPASSRnet/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":284988465,"owners_count":27095952,"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","status":"online","status_checked_at":"2025-11-18T02:00:05.759Z","response_time":61,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"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":["depth-estimation","parallax","stereo-matching","super-resolution"],"created_at":"2025-11-13T08:00:36.460Z","updated_at":"2025-11-18T02:02:06.190Z","avatar_url":"https://github.com/SYSU-SAIL.png","language":"Python","funding_links":[],"categories":["LowLevelVision"],"sub_categories":["3D SemanticSeg"],"readme":"# PASSRnet: Parallax Attention Stereo Super-Resolution Network\nPytorch implementation of \"Learning Parallax Attention for Stereo Image Super-Resolution\", CVPR 2019\n\n[[arXiv]](https://arxiv.org/abs/1903.05784) [[CVF]](http://openaccess.thecvf.com/content_CVPR_2019/papers/Wang_Learning_Parallax_Attention_for_Stereo_Image_Super-Resolution_CVPR_2019_paper.pdf)  [[Supp]](http://openaccess.thecvf.com/content_CVPR_2019/supplemental/Wang_Learning_Parallax_Attention_CVPR_2019_supplemental.pdf) \n\n## Overview\n![overview](./Figs/Overview.png)\n\nFigure 1. Overview of our PASSRnet network.\n\n\u003cimg width=\"400\" src=\"https://github.com/LongguangWang/PASSRnet/blob/master/Figs/Parallax-attention.png\"/\u003e\u003c/div\u003e\n\nFigure 2. Illustration of our parallax-attention mechanism. \n\n\u003cimg width=\"500\" src=\"https://github.com/LongguangWang/PASSRnet/blob/master/Figs/Toy-example.png\"/\u003e\u003c/div\u003e\n\nFigure 3. A toy example illustration of the parallax-attention and cycle-attention maps generated by our PAM.\nThe attention maps (30×30) correspond to the regions (1×30) marked by a yellow stroke. In (a) and (b), the first row\nrepresents left/right stereo images, the second row stands for parallax-attention maps, and the last row represents cycle-attention maps.\n\n## [Flickr1024 Dataset](https://yingqianwang.github.io/Flickr1024/)\n\n\u003cimg width=\"500\" src=\"https://github.com/LongguangWang/PASSRnet/blob/master/Figs/Flickr1024.jpg\"/\u003e\u003c/div\u003e\n\nFigure 4. The Flickr1024 dataset.\n\n## Requirements\n- pytorch (0.4), torchvision (0.2) (Note: The code is tested with `python=3.6, cuda=9.0`)\n- Matlab (For training/test data generation)\n\n## Train\n### Prepare training data\n1. Download the Flickr1024 dataset and put the images in `data/train/Flickr1024` \n(Note: In our paper, we also use 60 images in the Middlebury dataset as the training set.)\n2. Cd to `data/train` and run `generate_trainset.m` to generate training data.\n\n### Begin to train\n```bash\npython train.py --scale_factor 4 --device cuda:0 --batch_size 32 --n_epochs 80 --n_steps 30\n```\n\n## Test\n### Prepare test data\n1. Download the [KITTI2012](http://www.cvlibs.net/datasets/kitti/eval_stereo_flow.php?benchmark=stereo) dataset and put folders `testing/colored_0` and `testing/colored_1` in `data/test/KITTI2012/original` \n2. Cd to `data/test` and run `generate_testset.m` to generate test data.\n3. (optional) You can also download KITTI2015, Middlebury or other stereo datasets and prepare test data in `data/test` as below:\n```\n  data\n  └── test\n      ├── dataset_1\n            ├── hr\n                ├── scene_1\n                      ├── hr0.png\n                      └── hr1.png\n                ├── ...\n                └── scene_M\n            └── lr_x4\n                ├── scene_1\n                      ├── lr0.png\n                      └── lr1.png\n                ├── ...\n                └── scene_M\n      ├── ...\n      └── dataset_N\n```\n\n### Demo\n```bash\npython demo_test.py --scale_factor 4 --device cuda:0 --dataset KITTI2012\n```\n\n## Results\n![2x](./Figs/results_2x_KITTI2012_KITTI2015.png)\n\nFigure 5. Visual comparison for 2× SR. These results are achieved on “test_image_013” of the KITTI 2012 dataset and “test_image_019” of the KITTI 2015 dataset. \n\n![4x](./Figs/results_4x_KITTI2015.png)\n\nFigure 6. Visual comparison for 4× SR. These results are achieved on “test_image_004” of the KITTI 2015 dataset.\n\n![2x](./Figs/results_2x_lab.png)\n\nFigure 7. Visual comparison for 2× SR. These results are achieved on a stereo image pair acquired in our laboratory.\n\n## Citation\n```\n@InProceedings{Wang2019Learning,\n  author    = {Longguang Wang and Yingqian Wang and Zhengfa Liang and Zaiping Lin and Jungang Yang and Wei An and Yulan Guo},\n  title     = {Learning Parallax Attention for Stereo Image Super-Resolution},\n  booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},\n  year      = {2019},\n}\n```\n## Contact\nFor questions, please send an email to wanglongguang15@nudt.edu.cn\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FSYSU-SAIL%2FPASSRnet","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FSYSU-SAIL%2FPASSRnet","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FSYSU-SAIL%2FPASSRnet/lists"}