Retinex-Based Image Defogging with Three Algorithm Implementations
- Login to Download
- 1 Credits
Resource Overview
Image defogging code based on Retinex theory, implementing three distinct methods: Single-Scale Retinex (SSR), Multi-Scale Retinex (MSR), and Self-Quotient Image (SQI) approach. Execute the test files in respective folders to generate corresponding defogging results, featuring Gaussian filtering implementations and parameter optimization capabilities.
Detailed Documentation
This code provides Retinex theory-based image defogging implementations through three specialized algorithms. The implementation includes Single-Scale Retinex utilizing Gaussian filtering with adjustable sigma parameters, Multi-Scale Retinex combining multiple Gaussian scales through weighted summation, and Self-Quotient Image method that enhances illumination invariance. Each method is organized in separate folders containing test scripts that automatically load sample images, apply the algorithm, and display comparative results. Users can modify key parameters like Gaussian kernel sizes, scale weights, and normalization factors to optimize performance for specific fog conditions. The code structure facilitates easy experimentation with different Retinex variants while maintaining modular components for core operations like logarithmic transformation, convolution processing, and color restoration.
- Login to Download
- 1 Credits