GPU-CLAHE 🚀

Ultra-fast GPU-accelerated CLAHE implementation achieving 7000+ images/second

🔥 Performance Highlights

• 7,984 images/second on modern GPUs
• 80-160x faster than traditional CPU implementations
• 100% GPU utilization with zero CPU bottlenecks
• Memory efficient batch processing
• Production ready for million-image datasets

📦 Installation

pip install gpu-clahe

🚀 Quick Start

import gpu_clahe
import numpy as np

# Load your images (batch, height, width)
images = np.random.randint(0, 256, (1000, 512, 512), dtype=np.uint8)

# Process with GPU-CLAHE
result = gpu_clahe.process_images(images)

print(f"Processed {len(result)} images in seconds!")

💡 Advanced Usage

import gpu_clahe

# Custom configuration
config = gpu_clahe.CLAHEConfig(
    tile_size=16,
    clip_limit=3.0,
    enable_xla=True
)

# Process with custom settings
result = gpu_clahe.process_images(
    images,
    config=config,
    batch_size=64
)

# Benchmark performance
benchmark = gpu_clahe.benchmark_performance(
    image_shape=(1000, 512, 512)
)
print(f"Peak performance: {max([r['images_per_sec'] for r in benchmark['batch_results']]):.0f} img/sec")

📊 Benchmarks

Method	Images/Second	Relative Speed
OpenCV (CPU)	50-100	1x
GPU-CLAHE	7,984	80-160x

🎯 Use Cases

• Medical imaging pipelines
• Computer vision preprocessing
• Large-scale data augmentation
• Real-time video processing
• Batch image enhancement

📚 Documentation

Full documentation available at: gpu-clahe.readthedocs.io

🤝 Contributing

Contributions welcome! See CONTRIBUTING.md for guidelines.

📄 License

MIT License - see LICENSE file for details.

🙏 Citation

If you use this in research, please cite:

@software{gpu_clahe,
  author = {Baha2rM98},
  title = {GPU-CLAHE: Ultra-fast GPU-accelerated CLAHE implementation},
  url = {https://github.com/Baha2rM98/gpu-clahe},
  year = {2025}
}