concordia: sync reference capture, class figure and caption with GitHub

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

The reference capture moves from `image_1808` (point targets) to `image_0753` (diverse), so `assets/reference_capture.png` now shows the B-mode reconstruction next to `data/diverse_source_image` and the `data/scatterers` cloud on shared mm axes.

`assets/classes.png` is redrawn as a class-by-field grid: one capture per class in the columns (`image_0012`, `image_0410`, `image_0733`, `image_0753`, `image_1808`), with the reference B-mode (`data/image`), the scatterer cloud and the label map in the rows. The hero caption links the new captures and names the field behind each row.

**GitHub PR:**
- [tristan-deep/OpenH-RF#46](https://github.com/tristan-deep/OpenH-RF/pull/46) — concordia: choose reference frame, update assets and README

Files changed: 3.

```
~ concordia/README.md
~ concordia/assets/classes.png
~ concordia/assets/reference_capture.png
```

concordia/README.md

@@ -21,7 +21,7 @@ size_categories:
 
 ![B-mode reconstructions of the five SynthUS-FSA phantom classes](assets/classes.png)
 
-*The five phantom classes (anechoic, hypoechoic, hyperechoic, diverse, point-target), each reconstructed from `data/raw_data`: [`image_0005`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_0005.hdf5), [`image_0470`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_0470.hdf5), [`image_0640`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_0640.hdf5), [`image_1100`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_1100.hdf5), [`image_1808`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_1808.hdf5).*
+*The five phantom classes (anechoic, hypoechoic, hyperechoic, diverse, point-target), one capture per column: [`image_0012`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_0012.hdf5), [`image_0410`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_0410.hdf5), [`image_0733`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_0733.hdf5), [`image_0753`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_0753.hdf5), [`image_1808`](https://huggingface.co/datasets/nvidia/OpenH-RF/blob/main/concordia/data/image_1808.hdf5). Rows show the reference B-mode (`data/image`), the simulated scatterer cloud (`data/scatterers`) and the label map (`data/segmentation` for the lesion classes, `data/diverse_source_image` for diverse; point-target captures store none).*
 
 ## Dataset Description