{"id":21529797,"url":"https://github.com/guopengf/reconformer","last_synced_at":"2025-10-06T23:22:26.000Z","repository":{"id":62642814,"uuid":"449429855","full_name":"guopengf/ReconFormer","owner":"guopengf","description":"ReconFormer: Accelerated MRI Reconstruction Using Recurrent Transformer","archived":false,"fork":false,"pushed_at":"2024-08-03T11:37:11.000Z","size":109,"stargazers_count":65,"open_issues_count":8,"forks_count":10,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-04-23T13:37:50.892Z","etag":null,"topics":["deep-learning","mri-reconstruction","pytorch","transformer"],"latest_commit_sha":null,"homepage":"https://arxiv.org/abs/2201.09376","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/guopengf.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,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2022-01-18T20:03:19.000Z","updated_at":"2025-04-22T05:09:58.000Z","dependencies_parsed_at":"2025-01-09T19:41:48.661Z","dependency_job_id":"9927e0b4-5f50-4c24-ae26-d42ac8870191","html_url":"https://github.com/guopengf/ReconFormer","commit_stats":{"total_commits":20,"total_committers":2,"mean_commits":10.0,"dds":0.4,"last_synced_commit":"e2e0d5e6e58e04ad1c77a1151e63cf56bec21fb1"},"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/guopengf/ReconFormer","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/guopengf%2FReconFormer","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/guopengf%2FReconFormer/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/guopengf%2FReconFormer/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/guopengf%2FReconFormer/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/guopengf","download_url":"https://codeload.github.com/guopengf/ReconFormer/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/guopengf%2FReconFormer/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":278693807,"owners_count":26029601,"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":"2025-10-06T02:00:05.630Z","response_time":65,"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":["deep-learning","mri-reconstruction","pytorch","transformer"],"created_at":"2024-11-24T01:59:09.573Z","updated_at":"2025-10-06T23:22:25.966Z","avatar_url":"https://github.com/guopengf.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# ReconFormer\nReconFormer: Accelerated MRI Reconstruction Using Recurrent Transformer\n\nPytorch Code for the paper [\"ReconFormer: Accelerated MRI Reconstruction Using Recurrent Transformer\"](https://arxiv.org/abs/2201.09376)\n\n**Updates**:\n\n:rocket: : We release training and testing code \\\n:rocket: : We release pre-trained weights for fastMRI \n\n# Requirements\n\npython=3.6  \npytorch=1.7.0\n\nPlease refer conda_environment.yml for more dependencies.\n\n# Inroduction\n\nAccelerating magnetic resonance imaging (MRI) reconstruction process is a challenging ill-posed inverse problem due to the excessive under-sampling operation in k-space.\nIn this paper, we propose a recurrent transformer model, namely ReconFormer, for MRI reconstruction which can iteratively reconstruct high fidelity magnetic resonance images from highly under-sampled k-space data. In particular, the proposed architecture is built upon Recurrent Pyramid Transformer Layers (RPTLs), which jointly exploits intrinsic multi-scale information at every architecture unit as well as the dependencies of the deep feature correlation through recurrent states. Moreover, the proposed ReconFormer is lightweight since it employs the recurrent structure for its parameter efficiency.\n\n## Dataset Preparation\n\nPrepare the dataset in the following structure for easy use of the code.The provided data loaders is ready for this this format and you may change it as your need.\n\n```bash\n\n\n                   |-- \n                   |                       \n                   |                |--xxx.h5  \nDataset Folder-----|      |--train--|...\n                   |      |         |...\n                   |      |                  \n                   |      |         \n                   |--PD -|\n                          |\n                          |         |--xxx.h5 \n                          |-- val --|...  \n                                    |...\n ```\n\n## Links for downloading the public datasets:\n\n1) fastMRI Dataset - \u003ca href=\"https://fastmri.med.nyu.edu/\"\u003e Link \u003c/a\u003e\n2) HPKS Dataset - We don't obtain the permission from Johns Hopkins Hospital to release this dataset.\n\nPreprocessed fastMRI (OneDrive) - \u003ca href=\"https://livejohnshopkins-my.sharepoint.com/:f:/g/personal/pguo4_jh_edu/EtXsMeyrJB1Pn-JOjM_UqhUB9-QAehQs4cLwomJS2SkpGA?e=IUfPrp\"\u003e Link \u003c/a\u003e\\\nPassword: pguo4\\\n**Note:** In preprocessed fastMRI, We didn't modify the original fastMRI data and just make the format compatible with our DataLoader. \n\n# Run\n\n## Clone this repo\n```bash \ngit clone git@github.com:guopengf/ReconFormer.git\n```\n\n## Set up conda environment\n```bash\ncd ReconFormer\nconda env create -f conda_environment.yml\nconda activate recon\n```\n## Train ReconFormer\n```bash \nbash run_recon_exp.sh\n```\n\n## Monitor the traning process\n```bash \ntensorboard --logdir 'Dir path for saving checkpoints'\n```\n## Test \n(Download [pre-trained weights](https://livejohnshopkins-my.sharepoint.com/:f:/g/personal/pguo4_jh_edu/Er37oIyNy3NBrXbeCQBp_fQBAxELR8UDaq6gHd-fjwRrSw) Password: pguo4@jhu.edu)\n```bash \nbash run_recon_eval.sh\n```\n## Ackonwledgements\n\nWe give acknowledgements to [fastMRI](https://github.com/facebookresearch/fastMRI), [Swin-Transformer\n](https://github.com/microsoft/Swin-Transformer), and [SwinIR](https://github.com/JingyunLiang/SwinIR).\n\n\n# Citation\n```bash\n@article{guo2022reconformer,\n  title={ReconFormer: Accelerated MRI Reconstruction Using Recurrent Transformer},\n  author={Guo, Pengfei and Mei, Yiqun and Zhou, Jinyuan and Jiang, Shanshan and Patel, Vishal M},\n  journal={arXiv preprint arXiv:2201.09376},\n  year={2022}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fguopengf%2Freconformer","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fguopengf%2Freconformer","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fguopengf%2Freconformer/lists"}