README.md

---
tags:
- seismic
- deblending
- common-receiver
- segy
- geophysics
task_categories:
- image-to-image
- other
pretty_name: Deblending Common-Receiver Dataset
viewer: false
---

# Deblending Common-Receiver Dataset

Paired common-receiver SEG-Y volumes for supervised seismic deblending.

## Task

This dataset is used for direct paired regression:

```text
denoised = model(input)
```

Where:

- `input` is the pseudo-deblended common-receiver gather volume
- `target` is the cleaner forwarding-acoustic common-receiver gather volume

This is not a noise-label regression dataset.

## Dataset Description

- **Input variants currently included**:
- `T02_comp`: `input/T02_pseudo_deblended_common_receiver_comp.sgy`
- `T02_mod`: `input/T02_pseudo_deblended_common_receiver_mod.sgy`
- `T02_simp`: `input/T02_pseudo_deblended_common_receiver_simp.sgy`
- **Shared target file(s)**:
- `target/T02_forwarding_acoustic_common_receiver.sgy`
- **Geometry used by the benchmark**: `(386, 270, 2000)`
- **Interpretation**:
  - 386 common-receiver gathers
  - 270 traces per gather
  - 2000 time samples per trace
- **Format**: paired SEG-Y volumes with matching sequential gather layout

## Preview Images

<div align="center">

  <p><b>Clean Reference</b></p>
  <img src="assets/deblending/clean.png" alt="clean reference" width="92%">

  <p><b>Simple Blending Input</b></p>
  <img src="assets/deblending/simple.png" alt="simple blending input" width="92%">

  <p><b>Moderate Blending Input</b></p>
  <img src="assets/deblending/moderate.png" alt="moderate blending input" width="92%">

  <p><b>Hard Blending Input</b></p>
  <img src="assets/deblending/hard.png" alt="hard blending input" width="92%">

</div>


## File Structure

| Kind | Path | Variant | Size |
|------|------|---------|------|
| input | `input/T02_pseudo_deblended_common_receiver_comp.sgy` | `T02_comp` | 819.0 MB |
| input | `input/T02_pseudo_deblended_common_receiver_mod.sgy` | `T02_mod` | 819.0 MB |
| input | `input/T02_pseudo_deblended_common_receiver_simp.sgy` | `T02_simp` | 819.0 MB |
| target | `target/T02_forwarding_acoustic_common_receiver.sgy` | shared | 819.0 MB |

## Loading Data

```python
import numpy as np
import segyio


def read_common_receiver_volume(path, traces_per_gather=270):
    with segyio.open(path, "r", ignore_geometry=True) as src:
        n_traces_total = src.tracecount
        n_time = src.samples.size
        n_gathers = n_traces_total // traces_per_gather
        data = segyio.tools.collect(src.trace[:]).astype(np.float32, copy=False)
        data = data.reshape(n_gathers, traces_per_gather, n_time)
    return data


input_volume = read_common_receiver_volume(
    "input/T02_pseudo_deblended_common_receiver_mod.sgy",
    traces_per_gather=270,
)
target_volume = read_common_receiver_volume(
    "target/T02_forwarding_acoustic_common_receiver.sgy",
    traces_per_gather=270,
)
```

With `huggingface_hub`:

```python
from huggingface_hub import hf_hub_download

input_path = hf_hub_download(
    repo_id="GeoBrain/deblending-common-receiver",
    filename="input/T02_pseudo_deblended_common_receiver_mod.sgy",
    repo_type="dataset",
)
```

## Benchmark Split

The current benchmark uses a sequential gather split:

| Split | Gathers |
|-------|---------|
| Train | 300 |
| Val | 43 |
| Test | 43 |

The split is applied on common-receiver gather index.

## Preprocessing Recipe

The companion benchmark currently applies:

1. shared `max_abs` normalization between input and target
2. per-gather normalization through the shared `shot` normalization scope
3. overlapping 2D trace-time patches

Current patch settings:

- `patch_trace: 128`
- `patch_time: 256`
- `patch_overlap: 0.5`

## Notes

- The benchmark uses common-receiver gathers, not common-shot gathers.
- Input and target must keep identical ordering and identical shape.
- The current benchmark reads these files by fixed sequential grouping with `270` traces per gather.
- Multiple pseudo-deblended input variants can share the same forwarding-acoustic target.

## Citation

If you use this dataset, cite the dataset repository:

```bibtex
@misc{deblending_common_receiver_dataset,
  title={Deblending Common-Receiver Dataset},
  howpublished={https://huggingface.co/datasets/GeoBrain/deblending-common-receiver},
}
```

## References

- `segyio` library: https://github.com/equinor/segyio