🎨 AI Image Colorizer Pro

Professional AI-powered image colorization using multiple deep learning models

🌐 Live Demo · 🐛 Report Bug · 💬 Discussions

---

📖 Table of Contents

• ✨ Features
• 🎯 Demo
• 🧠 Model Zoo
• ⚙️ Processing Pipeline
• 📊 Performance Benchmarks
• 🖼️ Visual Comparison
• 🚀 Installation
• 🧪 Testing
• 📈 Benchmark
• 📄 License

---

✨ Features

Feature	Description
🎨 Multiple AI Models	SIGGRAPH17, ECCV16, Lightweight and Fast models
🖌 Interactive Color Hints	Guide the network with manual hint points
🤖 GPT Vision Analysis	Image understanding & intelligent color suggestions
🎛 Advanced Post Processing	Brightness, saturation, vibrance & contrast controls
📊 Quality Metrics	PSNR, SSIM and Colorfulness evaluation
🎥 Video Colorization	Frame-by-frame video processing (Beta)
📈 Benchmark Suite	Compare models on standard datasets

---

🎯 Demo

---

🧠 Model Zoo

Model	Quality	Speed	Best For
SIGGRAPH17	⭐⭐⭐⭐⭐	⭐⭐☆☆☆	Highest Quality
ECCV16	⭐⭐⭐⭐☆	⭐⭐⭐☆☆	Balanced Results
Lightweight	⭐⭐⭐☆☆	⭐⭐⭐⭐☆	Everyday Usage
Fast	⭐⭐☆☆☆	⭐⭐⭐⭐⭐	Real-time Preview

---

⚙️ Processing Pipeline

graph LR
  A[Upload Image] --> B[Select AI Model]
  B --> C[Optional Color Hints]
  C --> D[AI Colorization]
  D --> E[Post Processing]
  E --> F[Metrics]
  F --> G[Download]

Loading

---

📊 Performance Benchmarks

Model	PSNR ↑	SSIM ↑	LPIPS ↓	CPU	GPU
🥇 SIGGRAPH17	24.12	0.89	0.23	2.1 s	0.14 s
🥈 ECCV16	23.45	0.87	0.26	1.8 s	0.12 s
🥉 Lightweight	20.87	0.82	0.35	0.3 s	0.03 s
⚡ Fast	18.92	0.76	0.42	0.08 s	0.01 s

---

🖼️ Visual Comparison

Input	AI Output	Ground Truth

---

🚀 Installation

git clone https://github.com/yourusername/ai-image-colorizer.git
cd ai-image-colorizer
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
streamlit run app.py

---

🧪 Testing

pytest tests/

---

📈 Benchmark

python benchmark.py

---

📄 License

MIT License — see LICENSE

---

Made with ❤️ using
PyTorch • Streamlit • OpenAI • Hugging Face