🧠 AI CNN + YOLOv8 Waste Classification System

Deep Learning-based Image Classification & Object Detection System for Waste Materials

---

🚀 Project Overview

This project implements a hybrid AI system combining:

• 🧠 Convolutional Neural Networks (CNN) for image classification
• 🎯 YOLOv8 for object detection

The goal is to detect and classify waste materials from images using annotated datasets.

---

✨ Key Features

Feature	Description
🎯 Object Detection	YOLOv8 detects waste objects
🧠 Image Classification	CNN classifies waste types
📊 Training Pipeline	Custom training with epochs
🗂️ Dataset Handling	Annotated dataset support
⚙️ Configurable	YAML-based dataset config
📁 Modular Code	Separate scripts for tasks

---

🧱 System Architecture

Input Image
     ↓
YOLOv8 Detection Model
     ↓
Extracted Object Regions
     ↓
CNN Classification Model
     ↓
Predicted Waste Category

🛠️ Tech Stack

Component	Technology
Language	Python
Detection Model	YOLOv8
Classification Model	CNN
Data Handling	YAML / Custom Scripts
Training	PyTorch-based frameworks

📂 Project Structure
AI-CNN-Model---7-Classification/
│── dataset/images/        # Training images
│── runs/detect/train/     # YOLO training outputs
│── create_annotations.py  # Annotation generation script
│── data.yaml              # Dataset configuration
│── image_classification.py# CNN classification script
│── yolov8n.pt             # Pretrained YOLO model

⚙️ Installation & Setup
1. Clone Repository
git clone https://github.com/itsy-Wency/AI-CNN-Model---7-Classification.git
2. Install Dependencies
pip install ultralytics torch torchvision matplotlib opencv-python
3. Prepare Dataset
Place images inside:
dataset/images/
Ensure annotations are properly generated:
python create_annotations.py
4. Train YOLOv8 Model
yolo detect train data=data.yaml model=yolov8n.pt epochs=50
5. Run Classification
python image_classification.py

📊 Model Workflow
Dataset → Annotation → YOLO Training → Detection → CNN Classification → Output
🎯 Learning Outcomes

This project demonstrates:

Deep learning model integration (Detection + Classification)
Dataset annotation and preprocessing
YOLOv8 training pipeline
CNN-based classification logic
End-to-end ML workflow
🚀 Future Improvements
📱 Deploy as web/mobile app
🎥 Real-time video detection
📊 Model evaluation metrics dashboard
🔍 Multi-class waste categorization
☁️ Cloud deployment (Azure / AWS)

👨‍💻 Author
Wency Jorda
GitHub: https://github.com/itsy-Wency

⭐ Support

If you find this project useful, consider giving it a ⭐ on GitHub.

📜 License

This project is intended for educational and research purposes.