datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
OpenH-RF
OpenH-RF
Dataset Description
OpenH-RF is a community-driven dataset initiative building the open, shared foundation needed to train and evaluate AI models built on pre-beamformed (channel capture) medical ultrasound measurements.
This dataset is a collection of RF samples and metadata in the zea file format from a variety of tasks and applications, including ultrasound localization microscopy, ultrasound computer tomography, b-mode, flow imaging and more.
Each… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/OpenH-RF.OpenH-RF-eSAF
OpenH-RF Sub-Dataset — Rotational 3D US Raw Channel Data for Elevational SAF (Simulated + Measured Phantom)
One .hdf5 per acquisition, zea file format. Raw per-element channel data
(pre-beamforming). See data/manifest.json for the full list and reconstruct.py
pipeline.yaml for the reference reconstruction.
Paths in this card are relative to the package root (this file ships as the
package README.md; the .hdf5 live in data/). Files marked (source) are
part of the authors'… See the full description on the dataset page: https://huggingface.co/datasets/RyoMurakami/OpenH-RF-eSAF.OpenH-RF-eSAF-raw
OpenH-RF eSAF — Raw Source Data Archive
This is a provenance/reproducibility archive — not itself the OpenH-RF
submission. The submitted dataset (195 zea-format .hdf5 acquisitions + data card
reconstruction pipeline) is at
https://huggingface.co/datasets/RyoMurakami/OpenH-RF-eSAF. This archive
preserves the raw, upstream acquisition/simulation outputs that were converted into
that packaged submission — released under the same CC BY 4.0 license (see
LICENSE).
Code and full… See the full description on the dataset page: https://huggingface.co/datasets/RyoMurakami/OpenH-RF-eSAF-raw.