README.md

---
license: other
license_name: multi-source-aggregated-terms
license_link: LICENSE
pretty_name: "Person, Face & Package — Home Security Detection (YOLO)"
task_categories:
  - object-detection
tags:
  - image
  - object-detection
  - computer-vision
  - yolo
  - ultralytics
  - bbox
  - home-security
  - surveillance
  - face-detection
  - parcel-detection
  - delivery
  - vehicle-detection
  - pet-detection
  - bird-detection
annotations_creators:
  - crowdsourced
  - found
language:
  - en
---

# Person, face & package — home-security detection dataset

YOLO-format dataset for person, face, vehicles, small vehicles, parcels, pets, and birds in home / delivery / street scenes. Full data lives under `bigsplit/`; `sample/` is a 100-image preview subset (same layout: flat `images/` and `labels/`).

Training configs point at a `data.yaml` beside those folders. Both splits use the same class list (`nc` and `names`); only the root path and which image set is packaged differ. Labels are YOLO `.txt` files with one line per box: `class_id x_center y_center width height` (normalized 0–1).

## `data.yaml` — tags and layout

Class **tags** are the `names` list; order is the YOLO class index (`0` … `nc - 1`). Below is the canonical schema used by `sample/data.yaml` (checked in). The full split uses the same `nc` and `names` in `bigsplit/data.yaml`; `path`, `train`, and `val` resolve to that folder’s `images/` when you train from the dataset root.

**`sample/data.yaml`**

```yaml
# Preview sample (100 images); same layout as bigsplit — no split subfolders.
path: .
train: images
val: images
nc: 9
names:
  - person
  - face
  - vehicle
  - bike
  - bicycle
  - parcel
  - dog
  - cat
  - bird
```

When the full dataset is present, `bigsplit/data.yaml` uses the same `path` / `train` / `val` / `nc` / `names` (often with a file header comment describing the full split instead of the sample).

## Preview (bounding boxes)

Five PNGs are generated from the checked-in **`sample/`** images and YOLO labels. Each file stitches **three** frames **edge-to-edge** at equal height (**no padding** between panels). Boxes and class names are drawn from the corresponding `sample/labels/*.txt` files.

<p align="center">
  <img src="assets/preview_01.png" width="32%" alt="Sample preview 1 — three stitched frames" />
  <img src="assets/preview_02.png" width="32%" alt="Sample preview 2 — three stitched frames" />
  <img src="assets/preview_03.png" width="32%" alt="Sample preview 3 — three stitched frames" />
</p>
<p align="center">
  <img src="assets/preview_04.png" width="32%" alt="Sample preview 4 — three stitched frames" />
  <img src="assets/preview_05.png" width="32%" alt="Sample preview 5 — three stitched frames" />
</p>

[preview_01.png](assets/preview_01.png) · [preview_02.png](assets/preview_02.png) · [preview_03.png](assets/preview_03.png) · [preview_04.png](assets/preview_04.png) · [preview_05.png](assets/preview_05.png) · [preview_manifest.txt](assets/preview_manifest.txt) (source filenames per composite)

Regenerate after changing `sample/`:

```bash
python3 -m venv .venv   # once
.venv/bin/pip install pillow pyyaml
.venv/bin/python scripts/make_readme_preview.py
```

## Layout

- `bigsplit/data.yaml` — class list and paths (`train` / `val` both point at `images/`).
- `sample/data.yaml` — same schema for the preview subset.

## Classes (`nc: 9`)

| id | name     |
|---:|----------|
| 0 | person   |
| 1 | face     |
| 2 | vehicle  |
| 3 | bike     |
| 4 | bicycle  |
| 5 | parcel   |
| 6 | dog      |
| 7 | cat      |
| 8 | bird     |

## License

See `LICENSE`.