cnn-image-classifier

CNN Image Classifier - CIFAR-10

A modern Flask web application for image classification using a trained CIFAR-10 Convolutional Neural Network model.

Features

• Real CIFAR-10 Model: Uses a trained CNN model for accurate image classification
• OpenCV Integration: Optimized image processing with OpenCV for better performance
• 10 Object Classes: Classifies images into airplane, automobile, bird, cat, deer, dog, frog, horse, ship, and truck
• Image Preview: Real-time preview of uploaded images
• Confidence Scores: Shows prediction confidence with visual progress bars

Quickstart

1. Create and activate a virtual environment (Windows PowerShell):

python -m venv .venv
.venv\Scripts\Activate.ps1

1. Install dependencies:

pip install -r requirements.txt

1. Run the app:

python app.py

1. Open in browser: http://localhost:5000

Project Structure

CNN-Image-Classifier/
├── app.py                 # Main application entry point
├── app/
│   ├── __init__.py       # Flask app factory
│   ├── model.py          # CIFAR-10 model implementation
│   └── routes.py         # Flask routes and handlers
├── models/
│   └── model_cifar10.h5  # Trained CIFAR-10 model
├── templates/
│   ├── index.html        # Upload page with modern UI
│   └── result.html       # Results page with predictions
├── static/
│   └── css/
│       └── style.css     # Custom styles (legacy)
├── uploads/              # Temporary file storage
├── requirements.txt      # Python dependencies
├── .gitignore           # Git ignore rules
└── README.md            # This file

Dependencies

• Flask 3.0.3: Web framework
• TensorFlow 2.15.0: Deep learning framework for model loading
• OpenCV 4.8.1.78: Computer vision library for image processing
• NumPy 2.1.2: Numerical computing
• Pillow 10.4.0: Image processing
• Werkzeug 3.0.3: WSGI utilities

Model Information

The application uses a pre-trained CIFAR-10 CNN model that can classify images into 10 categories:

• Airplane: Aircraft and flying vehicles
• Automobile: Cars and road vehicles
• Bird: Various bird species
• Cat: Domestic and wild cats
• Deer: Deer and similar animals
• Dog: Dogs and canines
• Frog: Frogs and amphibians
• Horse: Horses and equines
• Ship: Boats and watercraft
• Truck: Trucks and large vehicles

Technical Details

Image Processing Pipeline

1. Upload: User uploads image through web interface
2. Preprocessing: OpenCV processes image (resize to 32x32, normalize to [0,1])
3. Prediction: CIFAR-10 model predicts class and confidence
4. Results: Beautiful results page displays prediction with confidence score

Performance Optimizations

• OpenCV Processing: Faster image operations compared to PIL
• Model Caching: Model loaded once and cached for subsequent requests
• Stream Processing: Direct file stream processing without temporary files
• Responsive Design: Optimized for all device sizes

Usage

1. Upload Image: Click the upload area or drag & drop an image
2. Preview: See a preview of your uploaded image
3. Predict: Click "Prediksi Gambar" to classify the image
4. View Results: See the prediction with confidence score and class information

Configuration

• Max Upload Size: 10 MB (configurable in app/__init__.py)
• Supported Formats: PNG, JPG, JPEG
• Model Path: models/model_cifar10.h5
• Image Size: Automatically resized to 32x32 pixels for CIFAR-10

Development

Adding New Models

To use a different model:

1. Replace models/model_cifar10.h5 with your trained model
2. Update class names in app/model.py
3. Adjust image preprocessing if needed
4. Update the UI to reflect new classes

Customizing the UI

The application uses Tailwind CSS for styling. Key files:

• templates/index.html: Upload interface
• templates/result.html: Results display
• Custom CSS in template <style> sections

Notes

• Uploaded files are temporarily stored in uploads/ directory
• Model files are stored in models/ directory for better organization
• The application uses Indonesian language for user interface
• All error messages are localized to Indonesian
• Professional design without emojis for business use

License

This project is for educational purposes. Please ensure you have proper licensing for any models or datasets used.