Multi-Object Detection using YOLO and Custom CNN

This project performs real-time multi-object detection and image classification using YOLOv8 and a custom-trained CNN on CIFAR-10. It captures webcam video, detects objects, classifies images, and displays results live.

Video Preview

Original Video Speed: 0.5X

Mini_Project_2.mp4

Features

• Real-time object detection using YOLOv8 segmentation
• Image classification using a custom CNN trained on CIFAR-10
• Live webcam video processing
• Model saving and loading

Requirements

Install dependencies with:

pip install -r requirements.txt

Usage

1. Clone the repository and navigate to the project folder.
2. Ensure you have a webcam connected.
3. Run the main script:

   python main.py

4. Follow the prompts to train the model and start detection.
5. Press 'q' to quit the video window.

Files

• main.py: Main script for training and detection
• yolov8n-seg.pt: YOLOv8 model weights
• image_classifier.keras: Saved CNN model
• requirements.txt: Python dependencies

Contributors

• Harith Kavish S
• Sharwan Krishnan P
• Sanjay R