PrathikshaR7/CycleGAN

Photo2Sketch CycleGAN

A deep learning project that converts photographs to artistic sketches using CycleGAN architecture. This project supports image upload, video processing, and live camera feed for real-time sketch generation.

Features

• —Photo to Sketch Conversion: Transform uploaded photos into artistic sketches
• —Video Processing: Convert entire videos to sketch animations
• —Live Camera Feed: Real-time sketch generation from webcam
• —Dual-Stage Processing: Enhanced quality with color transfer and post-processing
• —Firebase Authentication: Secure user authentication system
• —Responsive Web Interface: Modern UI built with Flask and Bootstrap

Quick Start

Local Development

1. 1.Clone the repository
2. 2.Install dependencies: pip install -r requirements.txt
3. 3.Run the application: python app.py HEAD
4. 4.Open http://localhost:7860 in your browser

1. 1.Open http://localhost:5000 in your browser 678f30c (Hugging_Face)

Docker Deployment

1. 1.Build the Docker image: docker build -t photo2sketch . HEAD
2. 2.Run the container: docker run -p 7860:7860 photo2sketch

1. 1.Run the container: docker run -p 5000:5000 photo2sketch 678f30c (Hugging_Face)

Model Architecture

This project uses CycleGAN with the following components:

• —Generator networks for photo→sketch and sketch→photo conversion
• —Discriminator networks for adversarial training
• —Color transfer pipeline for enhanced output quality
• —Post-processing for photorealistic results

Requirements

• —Python 3.10+
• —PyTorch 2.0+
• —OpenCV
• —Flask
• —Firebase Admin SDK
• —CUDA (optional, for GPU acceleration)

File Structure

├── app.py                 # Main Flask application
├── models/               # CycleGAN model definitions
├── checkpoints/          # Pre-trained model weights
├── static/              # Static assets (CSS, JS, images)
├── templates/           # HTML templates
├── uploads/             # User uploaded files
├── requirements.txt     # Python dependencies
└── Dockerfile          # Docker configuration

Usage

1. 1.Upload Photo: Select an image file and upload it for sketch conversion
2. 2.Process Video: Upload a video file to convert all frames to sketches
3. 3.Live Camera: Enable webcam for real-time sketch generation
4. 4.Download Results: Save generated sketches and processed media

Model Performance

The model achieves high-quality sketch generation with:

• —Structural preservation of original photos
• —Artistic sketch-style rendering
• —Fast inference speed suitable for real-time applications

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• —Based on CycleGAN architecture from Zhu et al.
• —Uses pre-trained models for photo-to-sketch conversion
• —Enhanced with color transfer and post-processing pipelines