waterloo-muscle: sync data card, pipeline grid and figures with GitHub

Syncs the waterloo-muscle card, pipeline and figures on the Hub with the reviewed state on GitHub (open-h/OpenH-RF).

The card counted frame-transmits as frames. The protocol yields 14,880 frame-transmits per participant, which is 480 stored frames; across 39 participants that is 18,720 stored frames and 580,320 frame-transmits, not "580K raw RF frames". The description and quantification sections now say both numbers explicitly.

`pipeline.yaml` gains the explicit grid and dynamic range the reference reconstruction uses: x +/-19 mm, z 0-80 mm on a 381 x 801 grid at -60 dB. Without it zea falls back to an aperture-wide, full-depth default.

Adds the cine loop the card now leads with, `assets/Acq_p35_Calf_left_calf_lateral_longitudinal_relaxed_pressure.gif`, together with the `zea process` one-liner that renders that acquisition straight from the Hub. `assets/reconstruct_output.png` is the reference still and `assets/main.png` the release image; neither is embedded in the card.

Files changed: 5.

```
+ waterloo-muscle/assets/Acq_p35_Calf_left_calf_lateral_longitudinal_relaxed_pressure.gif
+ waterloo-muscle/assets/main.png
+ waterloo-muscle/assets/reconstruct_output.png
~ waterloo-muscle/README.md
~ waterloo-muscle/pipeline.yaml
```

waterloo-muscle/README.md

@@ -20,11 +20,29 @@ size_categories:
 
 # UW-Muscle RF
 
+![Reconstructed cineloop from Acq_p35_Calf_left_calf_lateral_longitudinal_relaxed_pressure.hdf5](assets/Acq_p35_Calf_left_calf_lateral_longitudinal_relaxed_pressure.gif)
+
+Cine loop of [`Acq_p35_Calf_left_calf_lateral_longitudinal_relaxed_pressure.hdf5`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/waterloo-muscle/data/Acq_p35_Calf_left_calf_lateral_longitudinal_relaxed_pressure.hdf5), reconstructed from the raw
+channel data with the `pipeline.yaml` in this folder.
+
+`zea` renders it straight from the Hub:
+
+```bash
+zea process \
+  --dataset hf://nvidia/OpenH-RF/waterloo-muscle/data/Acq_p35_Calf_left_calf_lateral_longitudinal_relaxed_pressure.hdf5 \
+  --config hf://nvidia/OpenH-RF/waterloo-muscle/pipeline.yaml \
+  --n-frames 1 \
+  --save-as png
+```
+
+Swap `--n-frames 1 --save-as png` for `--save-as gif` to get the cine loop.
+
+
 Dataset consisting of raw RF data and speed of sound measurements acquired in an in vivo speed of sound study conducted by LITMUS @ University of Waterloo.
 
 ## Dataset Description
 
-This is a dataset consisting of 580K raw RF frames (plane wave) and associated SoS measurements using a programmable research scanner configured for high frame rate imaging. We used a rigorous image collection protocol based on landmarking according to bone markers to ensure image consistency. This data was collected as part of the following study:
+This is a dataset consisting of 18,720 raw RF frames (plane wave; 580,320 frame-transmits) and associated SoS measurements using a programmable research scanner configured for high frame rate imaging. We used a rigorous image collection protocol based on landmarking according to bone markers to ensure image consistency. This data was collected as part of the following study:
 
 D. Xiao, P. De La Torre, M. Saif El Nasr, A. J. Y. Chee, M. Mourtzakis, and A. C. H. Yu, “LivePulse-Echo Speed-of-Sound Estimation for Quality Assessment of Large Muscles in Humans,”Ultrasound in Medicine & Biology, vol. 51, no. 11, pp. 1925–1935, Nov. 2025.
 
@@ -96,7 +114,7 @@ All `coordinates` arrays are per-pixel Cartesian positions in metres, last axis
 
 **Current OpenH-RF release:** 1,248 HDF5 files; 576.19 GB (576,186,417,152 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.
 
-Data was collected from 39 participants, each with 32 unique images spanning calf/bicep/quad, axis and muscle state. For each imaging location, 15 frames were made per limb under minimal contact and with pressure. Given that each frame consisted of 31 steered plane waves, our protocol yielded a total of 15 repeats × 31 frames × 8 views × 2 sides × 2 pressure settings = 14880 raw RF frames per participant. In total, our dataset is expected to contain 580k frames of raw RF data.
+Data was collected from 39 participants, each with 32 unique images spanning calf/bicep/quad, axis and muscle state. For each imaging location, 15 frames were made per limb under minimal contact and with pressure. Given that each frame consisted of 31 steered plane waves, our protocol yielded a total of 15 repeats × 31 transmits × 8 views × 2 sides × 2 pressure settings = 14,880 frame-transmits per participant, i.e. 480 stored frames per participant. In total the dataset contains 18,720 stored frames and 580,320 frame-transmits (see the table below).
 
 All 1,248 HDF5 files are uploaded; current stored size and format version are reported above.
 

waterloo-muscle/pipeline.yaml

@@ -17,3 +17,15 @@ pipeline:
             - 0.0
             - 1.0
     - log_compress
+parameters:
+    xlims:
+    - -0.019
+    - 0.019
+    zlims:
+    - 0.0
+    - 0.08
+    grid_size_x: 381
+    grid_size_z: 801
+    dynamic_range:
+    - -60
+    - 0