🎨 Ancient Vision

Transform any image into authentic traditional art styles—including Japanese Ukiyo-e and Indian folk/temple art—using state-of-the-art AI technology.

Ancient Vision bridges the gap between modern photography and traditional art, allowing users to experience the timeless beauty of both Japanese and Indian heritage styles through cutting-edge artificial intelligence.

✨ Features

🎨 Artistic Transformation

• Authentic Traditional Styles: Transform images into Japanese Ukiyo-e and a wide range of Indian folk and temple art styles
• ControlNet Integration: Structural guidance while maintaining artistic style
  • 🎯 Canny Edge Detection: Preserves edges and outlines (automatically applied)
  • 🏔️ Depth Estimation: Maintains spatial relationships and depth
  • 🕺 Pose Detection: Preserves human poses and body structure
• Smart Prompting: Default prompt with optional user additions
  • Each style has a unique, culturally accurate prompt for best results
  • User prompts are appended to the default for enhanced results
• Optimized Parameters: Pre-configured settings for best results
  • Strength: 0.75
  • Guidance Scale: 8.5
  • Inference Steps: 25
  • Resolution: 512x512

Supported Art Styles

• Japanese Ukiyo-e (classic, landscape, portrait, nature, urban, seasonal)
• Indian Traditional Styles:
  • Gond Art: Tribal art with dot patterns, nature, and storytelling
  • Kalighat Painting: Bengali folk art, bold lines, mythological themes
  • Kangra Miniature: Himalayan miniature, delicate brushwork, court scenes
  • Kerala Mural: South Indian temple art, mythological scenes, vibrant colors
  • Madhubani Art: Bihar folk art, geometric designs, ritualistic motifs
  • Mandana Art: Rajasthani wall art, geometric and auspicious patterns
  • Pichwai Painting: Nathdwara temple art, Krishna themes, devotional motifs

📱 Modern Input Methods

• File Upload: Standard drag-and-drop or click-to-browse functionality
• Real-time Camera Capture:
  • 📸 Direct photo capture within the application
  • 🔄 Front/back camera switching for mobile devices
  • 🎯 High-definition capture (1280x720 resolution)
  • 📱 Mobile-first responsive design
  • 🛡️ Secure permission-based camera access

🚀 User Experience

• Intuitive Interface: Clean, minimalist design focused on the art creation process
• Instant Results: Fast processing with optimized AI pipeline
• Cross-platform: Works seamlessly on desktop and mobile browsers

🛠️ Tech Stack

Component	Technology	Purpose
Backend	Flask + PyTorch	AI model serving and API endpoints
Frontend	React + Material-UI	Modern, responsive user interface
AI Models	Stable Diffusion + Custom LoRA	Traditional art style transformation engine
ControlNet	Multiple ControlNet Models	Structural guidance and preservation
Image Processing	Canvas API + WebRTC	Real-time camera capture and processing

🚀 Quick Start

Automated Setup (Recommended)

# Clone the repository
git clone https://github.com/VinaySurwase/Ancient-Vision.git
cd Ancient-Vision

# Make the run script executable and start everything
chmod +x run.sh
./run.sh

Access Points

• Frontend Application: http://localhost:3000
• Backend API: http://localhost:5000
• API Documentation: http://localhost:5000/docs

📖 Usage Guide

🖼️ Image Input Methods

Traditional Upload

1. Click "Choose File" or drag and drop images into the upload area
2. Supported formats: JPEG, PNG, WebP (max 10MB)

Camera Capture

1. Click "Take Photo" button
2. Grant camera permissions when prompted by your browser
3. Use the live preview to frame your subject
4. Toggle between front/back cameras using the switch button
5. Click "Capture Photo" to take the picture
6. Your image is immediately ready for transformation

🎨 Transformation Workflow

1. Select Input: Upload a file or capture with camera
2. Choose Model: Select your art model (automatic detection of available models)
3. Add Context (Optional): Provide additional descriptive prompts that will be added to the default prompt
4. Configure Settings (Optional): Adjust advanced parameters if needed (strength, guidance scale, steps)
5. Transform: Click transform and let AI create your traditional artwork with automatic ControlNet guidance
6. Download: Save your traditional artwork

📋 Default Configuration

Parameter	Value	Purpose
Base Prompt	Each style has a unique, culturally accurate prompt (see below)	Core traditional style definition
Negative	Each style has a negative prompt to avoid modern/realistic elements	Prevents unwanted modern elements
Strength	0.75	Balance between original and transformed
Guidance	8.5	Optimal prompt adherence for style
Steps	25	Fast processing with quality results
Resolution	512x512	Optimized for model performance
ControlNet	Canny (0.8 strength)	Automatic edge preservation

📋 Camera Feature Requirements

Requirement	Details
Protocol	HTTPS required in production environments
Permissions	User must explicitly grant camera access
Connectivity	Active internet connection for AI processing
Performance	Modern device with adequate processing power

⚙️ Manual Setup

Prerequisites

• Python 3.8 or higher
• Node.js 16.0 or higher
• npm or yarn package manager

Backend Setup

# Install Python dependencies
pip install -r requirements.txt

# Start the Flask server
python app.py

Frontend Setup

# Navigate to frontend directory
cd frontend

# Install dependencies
npm install

# Start development server
npm start

Custom Models

# Add your custom models to the models directory
cp your_custom_model.safetensors models/
# Models are automatically detected and loaded

🔌 API Reference

Endpoints

Method	Endpoint	Description	Parameters
POST	/api/transform	Transform image to traditional style	image, prompt, style, control_type, etc
GET	/api/models	List available AI models	None
GET	/api/styles	List available art styles	None
GET	/api/controlnet	List available ControlNet options	None
GET	/api/health	Service health check	None

Example Request

curl -X POST http://localhost:5001/api/transform \
  -H "Content-Type: application/json" \
  -d '{
    "image": "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQ...",
    "prompt": "lotus flowers, peacocks, or mythological scene",
    "style": "gond",  # or kalighat, kangra, kerala_mural, madhubani, mandana, pichwai, classic_ukiyo, etc.
    "control_type": "canny",
    "controlnet_conditioning_scale": 1.0,
    "strength": 0.75,
    "guidance_scale": 8.5,
    "num_inference_steps": 25
  }'

👩‍💻 Development

Environment Setup

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install development dependencies
pip install -r requirements-dev.txt

Testing

# Frontend tests
cd frontend
npm test

# Backend tests
python -m pytest tests/

# Run specific test file
npm test -- ImageUpload.test.js

Code Quality

# Frontend linting
cd frontend
npm run lint

# Python code formatting
black .
flake8 .

🔒 Security & Privacy

Data Protection

• No Server Storage: Images are processed in memory and immediately discarded
• Local Processing: Camera captures remain on the client until transformation
• Secure Transmission: All API communications use HTTPS in production
• Privacy First: No user data collection or tracking

Camera Security

• Explicit Permissions: Camera access requires user consent
• Stream Management: Automatic cleanup of camera streams after use
• Browser Security: Leverages native browser security for camera access
• HTTPS Requirement: Camera functionality requires secure connections in production

🌐 Browser Compatibility

Browser	Version	Camera Support	WebRTC	Canvas API
Chrome	53+	✅ Full	✅	✅
Firefox	36+	✅ Full	✅	✅
Safari	11+	✅ Full	✅	✅
Edge	12+	✅ Full	✅	✅
Mobile Chrome	53+	✅ Full	✅	✅
Mobile Safari	11+	✅ Full	✅	✅

Feature Support Notes

• Camera switching may be limited on some older mobile devices
• HTTPS required for camera access in all modern browsers
• WebGL support recommended for optimal performance

🤝 Contributing

We welcome contributions from the community! Please follow these guidelines:

Getting Started

1. Fork the repository
2. Clone your fork locally
3. Create a feature branch (git checkout -b feature/amazing-feature)
4. Make your changes
5. Test thoroughly
6. Commit with clear messages (git commit -m 'Add amazing feature')
7. Push to your branch (git push origin feature/amazing-feature)
8. Submit a Pull Request

Code Standards

• Follow existing code style and conventions
• Add tests for new functionality
• Update documentation as needed
• Ensure all tests pass before submitting

📄 License

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

🙏 Acknowledgments

• Traditional Ukiyo-e artists who inspired this work
• The Stable Diffusion community for foundational AI models
• Contributors and testers who helped improve the application
• Japanese cultural preservation organizations

📞 Support

• Documentation: Check this README and inline code comments
• Issues: Use GitHub Issues for bug reports and feature requests
• Community: Join discussions in GitHub Discussions
• Updates: Watch the repository for latest releases and updates

---

⬆ Back to Top

Made with ❤️ for art enthusiasts and technology lovers