{"success":true,"data":{"author":"Tassilo Wald et al., Division of Medical Image Computing, German Cancer Research Center (DKFZ)","bucketAccessType":"public","bucketName":"MIC-DKFZ/OpenMind","bucketPrefix":"","category":"Neuroscience","cloudProvider":"huggingface","createdAt":"2026-11-17T19:29:32.669Z","description":"OpenMind is a pre-training corpus of 113,921 3D MRI volumes of the head and neck from 34,191 participants, pooled by the Division of Medical Image Computing at the German Cancer Research Center (DKFZ) from 800 OpenNeuro datasets. The card counts 23 MRI modalities and techniques (T1w and inplane T2 among them) acquired on more than 30 scanner models. Images follow a modified BIDS layout under OpenMind/ds_XXXXXX/sub-XX/anat/ as NIfTI (.nii.gz). Each image has a companion folder with a deface_mask.nii.gz marking anonymized regions and an fb_mask.nii.gz marking where anatomy is present, so reconstruction losses and contrastive crop sampling can skip defaced or empty voxels. openneuro_metadata.csv lists relative paths plus harmonized metadata: modality, scanner manufacturer, model and field strength for every image, and age, sex, handedness, BMI and health status where the source dataset reported them. The repository holds about 1 TB (measured 2026-11-17). The paper uses it to benchmark 3D self-supervised pre-training methods including MAE, SimMIM, S3D, SimCLR, Volume Fusion and SwinUNETR. Load files with nibabel or SimpleITK. The paper states the dataset is released under CC-BY-4.0; cite Wald et al., An OpenMind for 3D medical vision self-supervised learning (arXiv 2412.17041), and the source OpenNeuro datasets.","displayTitle":"OpenMind: 114k 3D head and neck MRI volumes pooled from 800 OpenNeuro datasets for self-supervised learning","huggingFaceUrl":"https://huggingface.co/datasets/MIC-DKFZ/OpenMind","id":"921295c4-5184-5859-85e7-346f59ad069c","license":"CC-BY-4.0","licenseEvidenceUrl":"https://arxiv.org/html/2412.17041","objectCountEstimate":100000,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2412.17041","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"MIC-DKFZ/OpenMind\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"MIC-DKFZ/OpenMind\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=MIC-DKFZ, license:cc-by-4.0, size_categories:n<1K, downloads=259570"],"format":"mixed","generatedAt":"2026-11-17T19:40:02.357Z","layout":{"example":"OpenMind/ds000001/sub-01/anat/sub-01_T1w.nii.gz","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","decoded":{"fields":{"dtype":"int16","ndim":3,"niftiVersion":1,"shape":"160,192,192","voxelSizesMm":"1,1.3333,1.3333"},"format":"nifti","summary":"NIfTI-1 [160, 192, 192] int16, 1,1.3333,1.3333 mm"},"encoding":"hex","key":"OpenMind/ds000001/sub-01/anat/sub-01_T1w.nii.gz","truncatedAt":2048}],"sampleObjects":["OpenMind/ds000001/sub-01/anat/sub-01_T1w.nii.gz","OpenMind/ds000001/sub-01/anat/sub-01_T1w__Data/deface_mask.nii.gz","OpenMind/ds000001/sub-01/anat/sub-01_T1w__Data/fb_mask.nii.gz","OpenMind/ds000001/sub-01/anat/sub-01_inplaneT2.nii.gz","OpenMind/ds000001/sub-01/anat/sub-01_inplaneT2__Data/deface_mask.nii.gz"],"sourceObjectKey":"OpenMind/ds000001/sub-01/anat/sub-01_T1w.nii.gz","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"MIC-DKFZ/OpenMind\", filename=\"OpenMind/ds000001/sub-01/anat/sub-01_T1w.nii.gz\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"OpenMind/ds000001/sub-01/anat/sub-01_T1w.nii.gz","totalObjects":100000},"summary":"OpenMind is a pre-training corpus of 113,921 3D MRI volumes of the head and neck from 34,191 participants, pooled by the Division of Medical Image Computing at the German Cancer…","title":"OpenMind: 114k 3D head and neck MRI volumes pooled from 800 OpenNeuro datasets for self-supervised learning","updatedAt":"2026-11-17T23:12:06.908Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/MIC-DKFZ/OpenMind"}}