dartmouth-uct: sync data card, pipeline and figures with GitHub
Browse filesSyncs the dartmouth-uct card, pipeline and figures on the Hub with the reviewed state on GitHub (open-h/OpenH-RF).
`pipeline.yaml` now calls `usct_reflectivity_das` bare. `compounding: coherent` is that operation's default, so the explicit parameter was redundant.
Adds the figure the card leads with, `assets/phantom_179604449_z200.png`: a 2D-sim breast cross-section reconstructed by `reconstruct.py` beside the ground-truth sound-speed and attenuation maps stored in the same file. `assets/main.png` is the release image and is not embedded in the card.
The card's links are fixed for the Hub: `reconstruct.py` now points at the script on GitHub rather than a path that only resolves in a checkout, the data file points at its blob here, and the references to `convert_2d_to_zea.py` / `convert_3d_to_zea.py` are dropped since neither is published. The run instructions match the script as it stands: set `ZEA_FILE` at the top and run it, with `SOS_MAP = True` swapping constant-sound-speed delays for a straight-ray integral through the ground-truth map.
Files changed: 4.
```
+ dartmouth-uct/assets/main.png
+ dartmouth-uct/assets/phantom_179604449_z200.png
~ dartmouth-uct/README.md
~ dartmouth-uct/pipeline.yaml
```
|
@@ -1,5 +1,12 @@
|
|
| 1 |
# OpenH-RF Sub-Dataset: 2D Ring-Array USCT Waveforms from 2D- and 3D-k-Wave Simulations
|
| 2 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
## Dataset Description
|
| 4 |
|
| 5 |
This sub-dataset provides **pre-beamformed channel-domain radio-frequency
|
|
@@ -102,10 +109,7 @@ plane**, y = elevation), and the voxel-level ground truth as native zea map fiel
|
|
| 102 |
per-pixel `coordinates`). Tissue class is stored in `metadata/annotations`
|
| 103 |
(`anatomy`, `label`); only fields with no standard spec home (z-plane indices,
|
| 104 |
element focus) are zea `CustomElement`s under the top-level `custom/` group. Files
|
| 105 |
-
are laid out by simulation type: `data/2d/` (2D-sim) and `data/3d/` (3D-sim)
|
| 106 |
-
produced by
|
| 107 |
-
[`convert_2d_to_zea.py`](convert_2d_to_zea.py) and
|
| 108 |
-
[`convert_3d_to_zea.py`](convert_3d_to_zea.py) respectively.
|
| 109 |
|
| 110 |
**Pre-processing applied before packaging:**
|
| 111 |
- 2D-sim: temporal decimation by 3× (20 MHz → 6.67 MHz native simulation rate).
|
|
@@ -197,7 +201,7 @@ Not applicable — all data are synthetic. Aggregate phantom statistics:
|
|
| 197 |
|
| 198 |
## Data Validation
|
| 199 |
|
| 200 |
-
A single reference reconstruction, [`reconstruct.py`](reconstruct.py), serves
|
| 201 |
**both** sub-datasets. It builds a `zea.Pipeline` whose beamforming stage is
|
| 202 |
zea's dedicated `zea.ops.USCTReflectivityDAS` — a round-trip time-of-flight
|
| 203 |
Delay-And-Sum that, for every pixel, coherently sums over all transmit/receive
|
|
@@ -214,10 +218,10 @@ drive the reconstruction directly. A resulting image whose bright skin boundary
|
|
| 214 |
traces the ground-truth contour confirms the geometry, timing, and transmit
|
| 215 |
parameters were recorded correctly.
|
| 216 |
|
| 217 |
-
```
|
| 218 |
-
|
| 219 |
-
|
| 220 |
-
|
| 221 |
|
| 222 |
## Known Issues
|
| 223 |
|
|
|
|
| 1 |
# OpenH-RF Sub-Dataset: 2D Ring-Array USCT Waveforms from 2D- and 3D-k-Wave Simulations
|
| 2 |
|
| 3 |
+

|
| 4 |
+
|
| 5 |
+
A 2D-sim breast cross-section,
|
| 6 |
+
[`data/2d/phantom_179604449_z200.hdf5`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/dartmouth-uct/data/2d/phantom_179604449_z200.hdf5),
|
| 7 |
+
reconstructed by [`reconstruct.py`](https://github.com/open-h/OpenH-RF/blob/main/datasets/dartmouth-uct/reconstruct.py) beside the ground-truth sound-speed
|
| 8 |
+
and attenuation maps stored in the file.
|
| 9 |
+
|
| 10 |
## Dataset Description
|
| 11 |
|
| 12 |
This sub-dataset provides **pre-beamformed channel-domain radio-frequency
|
|
|
|
| 109 |
per-pixel `coordinates`). Tissue class is stored in `metadata/annotations`
|
| 110 |
(`anatomy`, `label`); only fields with no standard spec home (z-plane indices,
|
| 111 |
element focus) are zea `CustomElement`s under the top-level `custom/` group. Files
|
| 112 |
+
are laid out by simulation type: `data/2d/` (2D-sim) and `data/3d/` (3D-sim).
|
|
|
|
|
|
|
|
|
|
| 113 |
|
| 114 |
**Pre-processing applied before packaging:**
|
| 115 |
- 2D-sim: temporal decimation by 3× (20 MHz → 6.67 MHz native simulation rate).
|
|
|
|
| 201 |
|
| 202 |
## Data Validation
|
| 203 |
|
| 204 |
+
A single reference reconstruction, [`reconstruct.py`](https://github.com/open-h/OpenH-RF/blob/main/datasets/dartmouth-uct/reconstruct.py), serves
|
| 205 |
**both** sub-datasets. It builds a `zea.Pipeline` whose beamforming stage is
|
| 206 |
zea's dedicated `zea.ops.USCTReflectivityDAS` — a round-trip time-of-flight
|
| 207 |
Delay-And-Sum that, for every pixel, coherently sums over all transmit/receive
|
|
|
|
| 218 |
traces the ground-truth contour confirms the geometry, timing, and transmit
|
| 219 |
parameters were recorded correctly.
|
| 220 |
|
| 221 |
+
Set `ZEA_FILE` at the top of the script and run `python reconstruct.py`; the
|
| 222 |
+
figure at the top of this card is its output. `SOS_MAP = True` swaps the
|
| 223 |
+
constant-sound-speed delays for a straight-ray integral through the
|
| 224 |
+
ground-truth map.
|
| 225 |
|
| 226 |
## Known Issues
|
| 227 |
|
|
Git LFS Details
|
|
Git LFS Details
|
|
@@ -5,9 +5,7 @@ pipeline:
|
|
| 5 |
dtype: float32
|
| 6 |
- name: patched_grid
|
| 7 |
operations:
|
| 8 |
-
-
|
| 9 |
-
params:
|
| 10 |
-
compounding: coherent
|
| 11 |
params:
|
| 12 |
with_batch_dim: false
|
| 13 |
jit_options: null
|
|
|
|
| 5 |
dtype: float32
|
| 6 |
- name: patched_grid
|
| 7 |
operations:
|
| 8 |
+
- usct_reflectivity_das
|
|
|
|
|
|
|
| 9 |
params:
|
| 10 |
with_batch_dim: false
|
| 11 |
jit_options: null
|