| --- |
| name: us4us |
| pretty_name: "us4us Ring-Array USCT (forearm, breast phantom, water)" |
| license: cc-by-4.0 |
| task_categories: |
| - other |
| tags: |
| - ultrasound |
| - rf |
| - open-h-rf |
| - zea |
| - usct |
| - tomography |
| - ring-array |
| - attenuation |
| - forearm |
| language: |
| - en |
| size_categories: |
| - n<1K |
| --- |
| |
| # us4us Ring-Array USCT |
|
|
|  |
|
|
| *A left-forearm cross-section, slice 8 of [`data/s1_l.hdf5`](https://huggingface.co/datasets/nvidia/Open-H-RF/blob/main/us4us/data/s1_l.hdf5), reconstructed beside the sinogram and reference image stored in the file.* |
|
|
| ## Dataset Description |
| Raw RF data acquired with the us4us Ltd. us4R system and a Draminski ring probe. |
|
|
| Intended for ultrasound tomographic reconstruction. |
|
|
| Acquisitions were performed on: |
| - the left and right human forearms of 8 healthy volunteers, |
| - a Yezitronix B-RG-1.2 breast phantom, |
| - a single reference slice from water only. |
|
|
| The forearm and phantom acquisitions comprise consecutive slices spaced 9 mm apart in depth (half of the probe elevation). |
|
|
| Each slice was recorded using 1024 subsequent single-element transmissions, and received by a 512-element aperture located opposite the transmitting element. |
|
|
| ## Dataset Contributor(s) |
|
|
| - Ziemowit Klimonda (PI) |
| - Piotr Jarosik |
| - Jakub Rozbicki |
| - Piotr Karwat |
| - Marcin Lewandowski |
| - us4us Ltd. (https://us4us.eu/) |
|
|
| ## Dataset Creation Date |
| 09/07/2026 |
|
|
| ## License / Terms of Use |
|
|
| [Creative Commons Attribution 4.0 International (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/legalcode.en). Retain attribution and identify modifications when reusing the data. |
|
|
| ## Intended Usage |
| Ultrasound computed tomography image reconstruction. |
|
|
| ## Dataset Characterization |
| * Data Collection Method: phantom, healthy adult human volunteers. |
| * Labeling Method: N/A. |
| * Acquisition system: |
| - ring probe: |
| - probe radius: 130mm, |
| - number of elements: 1024, |
| - center frequency: 2MHz, |
| - sampling rate: 8125000.0, |
| - us4R research system + host PC, |
| - custom positioning system for subject spatial control. |
| * Tx/Rx scheme: |
| - the scheme consisted of 1024 transmit-receive events, |
| - each event consisted of a single probe element transmitting a short pulse (1 period at 2 MHz excitation), while 512 probe elements were used for reception, |
| - the 1024 transmit events were performed sequentially, with each probe element transmitting once, |
| - for each transmission, the center of the receiving aperture was positioned on the opposite side of the probe relative to the transmitting element, |
| - for example, for the 0th transmission, the 0th probe element transmitted, and the receiving aperture consisted of probe elements 256–767. |
|
|
| ## Processing the Dataset |
|
|
| The acquisitions can be processed with the `reconstruct.py` [script](https://github.com/open-h/Open-H-RF/blob/main/datasets/us4us/reconstruct.py) as provided in the [Open-H-RF GitHub repository](https://github.com/open-h/Open-H-RF), together with the `pipeline.yaml` definition in this folder and the [zea library](https://github.com/tue-bmd/zea). The script streams the data from the Hugging Face Hub. |
|
|
| The script rebuilds the attenuation sinogram and filtered-backprojection image from `data/raw_data` and plots them beside the versions stored in the file. Set `ZEA_FILE` and `FRAME` (the slice) at the top of the script. |
|
|
| ## Dataset Format |
|
|
| [zea v0.1.6](https://github.com/tue-bmd/zea) |
|
|
| All sub-datasets are provided in the ZEA file format. No preprocessing was performed on the raw channel data. |
|
|
| ## Dataset Quantification |
|
|
| **Current Open-H-RF release:** 18 HDF5 files; 32.66 GB (32,658,423,808 bytes) stored; root `zea_version` **0.1.6**. Sizes include all HDF5 contents and use decimal units (MB = 10^6 bytes, GB = 10^9 bytes, TB = 10^12 bytes), not decoded-array memory or original-source download sizes. |
|
|
| * Each forearm HDF5 file contains 10 slices of the subject; `yezitronix-b-rg-1.2.hdf5` contains 11 and `reference_water.hdf5` a single slice. |
| * Stored HDF5 file sizes vary: 1.89 GB to 1.95 GB excluding the water reference. `reference_water.hdf5` is 183,107,584 bytes (183.11 MB). |
|
|
| ## Subject Metadata |
| The dataset contains data from: |
| * breast phantom Yezitronix B-RG-1.2, |
| * both forearms of 8 healthy human volunteers (2 females and 6 males, age range [20, 55]), |
| * water only (reference data). |
|
|
| ## Data Validation |
|
|
| * The data files contain attenuation sinograms and images reconstructed from raw data using the filtered backprojection algorithm. |
| * `reconstruct.py` rebuilds both from `data/raw_data` and plots them beside the stored versions (see *Processing the Dataset*). |
|
|
| ## Known Issues |
| * The following probe elements should be considered damaged: [367, 409, 457, 764, 775, 776, 368, 390, 470, 739, 792]. |
|
|
| ## Ethical Considerations |
| * All human subjects were healthy adults who voluntarily participated in the measurements. |
|
|
