README.md

---
license: mit
task_categories:
- object-detection
- image-classification
tags:
- animal-detection
- animal-reidentification
- bounding-box
- face-detection
- body-detection
- tracklets
- re-identification
size_categories:
- 1K<n<10K
configs:
- config_name: face_and_body
  data_files:
  - split: test
    path: face_and_body/test-*
- config_name: body_only
  data_files:
  - split: test
    path: body_only/test-*
- config_name: original_with_face_body_bbox
  data_files:
  - split: test
    path: original_with_face_body_bbox/test-*
- config_name: original_with_body_bbox
  data_files:
  - split: test
    path: original_with_body_bbox/test-*
- config_name: face_tracklets
  data_files:
  - split: test
    path: face_tracklets/test-*
- config_name: body_tracklets
  data_files:
  - split: test
    path: body_tracklets/test-*
---

# Zoo Animal Re-Identification Dataset

A dataset for animal re-identification with **2,705 body images**, **1,192 face crops**, and **6 configurations**.

## Configurations

### 1. `face_and_body`
Individual frames with both face and body crops.

**Features:**
- `date`: Date of capture (YYYY-MM-DD)
- `time`: Time of capture (HH:MM:SS)
- `class`: Animal name
- `video`: Source video filename
- `frame_number`: Frame number in video
- `camera`: Camera ID
- `face_image`: Cropped face image
- `body_image`: Cropped body/full image

### 2. `body_only`
Individual frames with body crops only.

**Features:**
- `date`, `time`, `class`, `video`, `frame_number`, `camera`: Same as above
- `body_image`: Cropped body image

### 3. `original_with_face_body_bbox`
Full original frames.

**Features:**
- `date`, `time`, `class`, `video`, `frame_number`, `camera`: Same as above
- `image`: Original full image

### 4. `original_with_body_bbox`
Full original frames.

**Features:**
- `date`, `time`, `class`, `video`, `frame_number`, `camera`: Same as above
- `image`: Original full image

### 5. `face_tracklets`
**Face images grouped by tracklet (animal + video)**. Each row contains all face crops from one animal in one video, ordered by frame number.

**Features:**
- `class`: Animal name
- `video`: Source video filename
- `camera`: Camera ID
- `frame_numbers`: List of frame numbers (ordered)
- `face_images`: Sequence of face images (ordered by frame)

### 6. `body_tracklets`
**Body images grouped by tracklet (animal + video)**. Each row contains all body crops from one animal in one video, ordered by frame number.

**Features:**
- `class`: Animal name
- `video`: Source video filename
- `camera`: Camera ID
- `frame_numbers`: List of frame numbers (ordered)
- `body_images`: Sequence of body images (ordered by frame)

## Usage

```python
from datasets import load_dataset

# Load individual frames with face and body
ds = load_dataset("Maxscha/test", "face_and_body", split="test")
print(ds[0]["face_image"])  # PIL Image
print(ds[0]["class"])  # Animal name

# Load face tracklets
ds = load_dataset("Maxscha/test", "face_tracklets", split="test")
print(len(ds[0]["face_images"]))  # Number of faces in this tracklet
print(ds[0]["class"])  # Animal name for this tracklet

# Load body tracklets
ds = load_dataset("Maxscha/test", "body_tracklets", split="test")
for img in ds[0]["body_images"]:
    print(img)  # Each PIL Image in the tracklet sequence
```

## Animals

The dataset contains images of 5 animals:
- Sango
- Tilla
- M'Penzi
- Bibi
- Djambala

## License

MIT