Fast Image Enhancement Based on Maximum and Guided Filters
Dan Zhu, Guannan Chen, Pablo Navarrete Michelini, Hanwen Liu · 2019
Retinex theory gives an effective tool for adjusting the brightness of low-light image by factorizing its values in terms of illumination and reflectance. In this paper, a Retinex image enhancement algorithm is proposed based on maximum and guided filters to improve the brightness of low-light images. The illumination is calculated by the maximum filter and edge-preserving smoothed by the guided filter. Both subjective and objective tests show that our method can achieve better visual quality than other state-of-the-art. Furthermore, because of the high computational efficiency of the maximum and guided filters, our method runs more than 7× faster than other state-of-the-art.