{"id":50713265,"url":"https://github.com/zekikus/medsegbench","last_synced_at":"2026-06-09T17:00:46.379Z","repository":{"id":254238503,"uuid":"845909555","full_name":"zekikus/MedSegBench","owner":"zekikus","description":"[pip install medsegbench] 35x Standardized Medical Segmentation Datasets from Different Data Modalities","archived":false,"fork":false,"pushed_at":"2024-11-27T11:06:58.000Z","size":209,"stargazers_count":25,"open_issues_count":0,"forks_count":2,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-09-23T14:06:09.135Z","etag":null,"topics":["2d","benchmark","benchmarking","deep-learning","medical-image-segmentation","medical-imaging","medsegbench","pytorch","segmentation","universal-segmentation"],"latest_commit_sha":null,"homepage":"https://medsegbench.github.io/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/zekikus.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":"2024-08-22T06:56:25.000Z","updated_at":"2025-09-09T13:05:05.000Z","dependencies_parsed_at":"2024-09-12T01:14:54.050Z","dependency_job_id":"c97fbb0d-3fe4-4072-bde5-92e30c7eb4b2","html_url":"https://github.com/zekikus/MedSegBench","commit_stats":null,"previous_names":["zekikus/medsegbench"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/zekikus/MedSegBench","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zekikus%2FMedSegBench","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zekikus%2FMedSegBench/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zekikus%2FMedSegBench/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zekikus%2FMedSegBench/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/zekikus","download_url":"https://codeload.github.com/zekikus/MedSegBench/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/zekikus%2FMedSegBench/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34116461,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-09T02:00:06.510Z","response_time":63,"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":["2d","benchmark","benchmarking","deep-learning","medical-image-segmentation","medical-imaging","medsegbench","pytorch","segmentation","universal-segmentation"],"created_at":"2026-06-09T17:00:30.635Z","updated_at":"2026-06-09T17:00:46.369Z","avatar_url":"https://github.com/zekikus.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# MedSegBench: A Comprehensive Benchmark for Medical Image Segmentation in Diverse Data Modalities\n## Data ([Zenodo](https://zenodo.org/records/13359660)) | Experiments ([Zenodo](https://zenodo.org/records/13381081)) | Paper ([Nature Scientific Data](https://www.nature.com/articles/s41597-024-04159-2)) | Website ([Website](https://medsegbench.github.io/))\n\n\n[Musa Aydin](https://scholar.google.com/citations?user=yfKMO-wAAAAJ\u0026hl=tr\u0026oi=ao), [Zeki Kuş](https://scholar.google.com/citations?user=h2B-3LwAAAAJ\u0026hl=tr\u0026oi=ao)\n\nMedSegBench is a comprehensive benchmark designed to evaluate deep learning models for medical image segmentation across a wide range of modalities. It covers a wide range of modalities, including 35 datasets with over 60,000 images from ultrasound, MRI, and X-ray. The benchmark addresses challenges in medical imaging by providing standardized datasets with train/validation/test splits, considering variability in image quality and dataset imbalances. The benchmark supports binary and multi-class segmentation tasks with up to 19 classes. It supports binary and multi-class segmentation tasks with up to 19 classes and uses the U-Net architecture with various encoder/decoder networks such as ResNets, EfficientNet, and DenseNet for evaluations. MedSegBench is a valuable resource for developing robust and flexible segmentation algorithms and allows for fair comparisons across different models, promoting the development of universal models for medical tasks. It is the most comprehensive study among medical segmentation datasets. The datasets and source code are publicly available, encouraging further research and development in medical image analysis.\n\n# Installation and Requirements\nSetup the required environments and install `medsegbench` as a standard Python package from [PyPI]([https://pypi.org/project/medmnist/](https://pypi.org/project/medsegbench/)):\n\n    pip install medsegbench\n\nOr install from source:\n\n    pip install --upgrade git+https://github.com/zekikus/MedSegBench.git\n\nThe code requires only common Python environments for machine learning. Basically, it was tested with\n* Python 3 (\u003e=3.6)\n* PyTorch\\==1.3.1\n* numpy\\==1.18.5, scikit-learn\\==0.22.2, Pillow\\==8.0.1\n* torchvision\\==0.11.2\n* segmentation-models-pytorch\\==0.3.3\n\n# Getting Started\nYou can use default 512 sized version using the downloaded files:\n\n    \u003e\u003e\u003e from medsegbench import Promise12MSBench\n    \u003e\u003e\u003e train_dataset = Promise12MSBench(split=\"train\")\n\nYou can download the dataset by setting `download=True`:\n\n    \u003e\u003e\u003e from medsegbench import Promise12MSBench\n    \u003e\u003e\u003e train_dataset = Promise12MSBench(split=\"train\", download=True)\n\nYou can download different sized versions by setting `size={128, 256, 512}`:\n\n    \u003e\u003e\u003e from medsegbench import Promise12MSBench\n    \u003e\u003e\u003e train_dataset = Promise12MSBench(split=\"train\", size=256)\n\nYou can download sub-categories of dataset by setting `category={C1, C2, C3, ...}`:\n\n    \u003e\u003e\u003e from medsegbench import WbcMSBench\n    \u003e\u003e\u003e train_dataset = WbcMSBench(split=\"train\", category='C1')\n\n## For Pytorch users\n* Code completely designed using Pytorch\n* Explore train and test scripts for MedSegBench. Using the following code files, you can reproduce reported results with shared weight in [Zenodo](https://zenodo.org/records/13381081)\n    * For Training:\n        * Binary Tasks: [`train_pytorch.py`](examples/train_pytorch.py)\n        * Multi-class Tasks: [`train_pytorch_multi_class.py`](examples/train_pytorch_multi_class.py)\n    * For Testing:\n        * Binary Tasks: [`test_pytorch.py`](examples/test_pytorch.py)\n        * Multi-class Tasks: [`test_pytorch_multi_class.py`](examples/test_pytorch_multi_class.py)\n          \n# Reproducibility\nYou can download trained model weights and detailed prediction results for each dataset from [Zenodo](https://zenodo.org/records/13381081)\n\n`dataset_name.zip:` It contains trained model weights for each encoder/decoder network with 3 different seed.\n\n`csv_files_predictions.zip:`\n\n* It includes detailed image-based prediction results for each dataset and model.\n* Each file named as {datasetname}\\_test\\_{modelname}\\_seed\\_{seedno}.csv\n\nAll performance measure scores for each dataset by seed: [`additional_files:`](additional_files/all_result.xlsx)\n\n# Citation\n\n    @article{Ku2024,\n      title = {MedSegBench: A comprehensive benchmark for medical image segmentation in diverse data modalities},\n      volume = {11},\n      ISSN = {2052-4463},\n      url = {http://dx.doi.org/10.1038/s41597-024-04159-2},\n      DOI = {10.1038/s41597-024-04159-2},\n      number = {1},\n      journal = {Scientific Data},\n      publisher = {Springer Science and Business Media LLC},\n      author = {Kuş,  Zeki and Aydin,  Musa},\n      year = {2024},\n      month = nov \n    }\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzekikus%2Fmedsegbench","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fzekikus%2Fmedsegbench","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzekikus%2Fmedsegbench/lists"}