Datasets:

---

nvidia

/

OpenH-RF

Like 61

Follow

NVIDIA 70.2k

Syncs the weillcornell card on the Hub with the reviewed state on GitHub (open-h/OpenH-RF).

The card now follows the layout shared by every OpenH-RF dataset: frontmatter with a name, a stand-alone title, the hero image with an italic caption linking the sample it shows, Description, Contributor(s), Creation Date and the standard CC BY 4.0 text, then a new Processing the Dataset section (the zea process command where the pipeline carries the whole reconstruction, and the reconstruct.py link) followed by Dataset Format, which opens with the zea version. Hard line breaks are gone so the Hub wraps paragraphs itself; the longer sections keep their content.

The card is rebuilt from its own sections into the common layout (Summary → Characterization, Suggested Tasks → Intended Usage, HDF5 Contents → Dataset Format, Limitations → Known Issues); the reference reconstruction becomes the hero.

GitHub PRs:

• tristan-deep/OpenH-RF#28 — Restructure all data cards to the common README layout

Files changed: 1.

~ weillcornell/README.md

Update — card tidy-up. The restructured card still carried content that was wrong here or not specific to this dataset:

• The Files section listed a LICENSE that does not exist, the reference figure at the folder root (it lives in assets/), and reconstruct.py (GitHub-only). It is removed; the scan-ID naming moves to Dataset Quantification.
• The hero figure is ac10_15m_SK frame 0 (its own title says so), but the caption and file name said ac1_15m_SK. The asset is renamed to assets/ac10_15m_SK_pipeline.png and the caption now links data/ac10_15m_SK.hdf5.
• Data Validation cited zea 0.1.3 validation and compression; the release is zea 0.1.6 (two files re-checked with the spec validator at 0.1.6: compliant).
• Generic OpenH-RF/zea text is dropped (environment setup, the Map attribute list, omitted annotations, "public reference targets").
• Characterization uses the labelled bullets, counts and the field table sit under Quantification, and a Subject Metadata section is added. Processing leads with the zea process one-liner, as on the other cards.

GitHub PR:

• tristan-deep/OpenH-RF#33 — weillcornell: tidy the data card to the common layout

Files changed: 3.

+ weillcornell/assets/ac10_15m_SK_pipeline.png
~ weillcornell/README.md
- weillcornell/assets/ac1_15m_SK_pipeline.png

Update — contributor figures and isotropic reconstruction grid. Brings in the Weill Cornell team's own card update (tristan-deep/OpenH-RF#47, merged on top of #33):

• The hero is now the team's 4-panel figure (assets/hero.png: B-mode, pooled Nakagami m and ω, theoretical BSC curves); assets/main.png adds their standalone B-mode.
• pipeline.yaml grid 300 × 597 → 600 × 301 (isotropic ~0.073 mm, ≈ λ/4). The old grid had pixels ~4× taller than wide, so zea process wrote a vertically stretched PNG; it now keeps the 2:1 physical aspect. The reference reconstruction is regenerated on the new grid and Data Validation cites it.

GitHub PR:

• tristan-deep/OpenH-RF#47 — weillcornell: update README and add asset images

Files changed: 5.

+ weillcornell/assets/hero.png
+ weillcornell/assets/main.png
~ weillcornell/README.md
~ weillcornell/pipeline.yaml
~ weillcornell/assets/ac10_15m_SK_pipeline.png