BFORE: Hybrid Metaheuristic Optimizes Retinex for Low-Light Images
A recent publication on arXiv (2605.03509) presents BFORE (Butterfly-Firefly Optimized Retinex Enhancement), a novel framework designed to automatically adjust parameters in a multi-stage Retinex-based system for enhancing low-light images. This technique transforms images into the HSV color space and utilizes Adaptive Gamma Correction with Weighted Distribution (AGCWD) on the luminance channel, followed by adaptive denoising. The optimization of Multi-Scale Retinex with Color Restoration (MSRCR) parameters is achieved through the Butterfly Optimization Algorithm (BOA), while the Firefly Algorithm (FA) fine-tunes AGCWD and denoising parameters. This innovative method overcomes the shortcomings of traditional Retinex techniques that depend on manually set parameters, which often struggle to adapt to varying lighting conditions.
Key facts
- arXiv paper 2605.03509 proposes BFORE
- BFORE stands for Butterfly-Firefly Optimized Retinex Enhancement
- Method uses HSV color space conversion
- Applies Adaptive Gamma Correction with Weighted Distribution (AGCWD)
- Butterfly Optimization Algorithm (BOA) optimizes MSRCR parameters
- Firefly Algorithm (FA) optimizes AGCWD and denoising parameters
- Addresses poor visibility, low contrast, and color distortion in low-light images
- Existing Retinex methods rely on manually tuned parameters
Entities
Institutions
- arXiv