1552444F

Chicken & Rabbit Detector 🐔🐰

An object detection application that identifies chickens and rabbits in images using a custom fine-tuned YOLOv11s model.

Features

• Image Detection: Upload an image to detect chickens and rabbits
• Video Detection: Process videos frame-by-frame with bounding box annotations to detect chicken and rabbit
• Bounding Box Visualization: Detected objects are highlighted with labeled bounding boxes
• Fast Inference: Optimized for efficient inference on uploaded images

Model Performance

• Architecture: YOLOv11s
• Training Strategy: Transfer learning with fine-tuning
• Early Stopping Epoch: 162
• mAP50: 96.2%
• mAP50:95: 89.1%
• Classes: Chicken, Rabbit

How to Use

Image Detection

1. Click on the "Image" tab
2. Upload an image containing chickens and/or rabbits
3. Click the "Detect" button
4. View the results with bounding boxes

Video Detection

1. Click on the "Video" tab
2. Upload a video containing chickens and/or rabbits
3. Click the "Detect" button
4. View the processed video with detections

Dataset

• Custom dataset manually annotated using Roboflow
• Includes realistic farm environments with varying poses, scales, and lighting
• Dataset link: https://universe.roboflow.com/goldfish-iiyzd/chicken_rabbit_detection/dataset/5

Project Information

• Student ID: 1552444F
• Course: ITI121 – Applied Deep Learning
• Institution: Nanyang Polytechnic