|
|
--- |
|
|
license: mit |
|
|
language: |
|
|
- en |
|
|
metrics: |
|
|
- precision |
|
|
- recall |
|
|
pipeline_tag: object-detection |
|
|
tags: |
|
|
- wild |
|
|
- life |
|
|
- animals |
|
|
- object |
|
|
- detection |
|
|
--- |
|
|
Wildlife Detection |
|
|
Author: Darshan Modi |
|
|
|
|
|
This model is a high‑performance object detection system trained on a curated dataset of African wildlife, including: |
|
|
- Buffalo |
|
|
- Elephant |
|
|
- Rhinoceros |
|
|
- Zebra |
|
|
The dataset contains diverse lighting conditions, camera angles, and natural environments, making the model suitable for wildlife monitoring, conservation research, anti‑poaching systems, and ecological analytics. |
|
|
|
|
|
🧠 Model Overview |
|
|
This ONNX model is optimized for real‑time inference on edge devices and cloud environments. It is built using the Ultralytics YOLO architecture and exported for ONNX runtime compatibility. |
|
|
Key Features |
|
|
- Detects multiple African wildlife species |
|
|
- Lightweight and fast inference |
|
|
- Suitable for Raspberry Pi, Jetson Nano, cloud servers, and desktop GPUs |
|
|
- Ideal for conservation AI, camera‑trap automation, and wildlife analytics |
|
|
|
|
|
📊 Performance Metrics |
|
|
Evaluated on a held‑out validation set from the same dataset: |
|
|
- Precision: 93% |
|
|
- Recall: 88% |
|
|
- mAP50: 94% |
|
|
These metrics indicate strong detection accuracy with low false positives and reliable species identification. |
|
|
|
|
|
📦 Intended Use |
|
|
This model is designed for: |
|
|
- Wildlife conservation projects |
|
|
- Camera‑trap automation |
|
|
- Animal population monitoring |
|
|
- Anti‑poaching surveillance |
|
|
- Ecological research |
|
|
- Real‑time detection on edge devices |
|
|
|
|
|
⚙️ Technical Details |
|
|
- Format: ONNX |
|
|
- Input size: 640×640 |
|
|
- Architecture: YOLO‑based detector |
|
|
- Classes: Buffalo, Elephant, Rhino, Zebra |
|
|
- Training epochs: 30 |
|
|
|
|
|
👤 Author |
|
|
Created by: Darshan Modi |
|
|
Focused on building practical AI systems for wildlife monitoring, safety applications, and real‑world deployment on edge devices. |
