Model for Single Image Enhancement Based on Adaptive Contrast Stretch and Multiband Fusion
Gengchen Xu, Chao Zheng, Yilin Gu · 2022 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS) · 2022
This paper proposes a rapid single image enhancement that may be used in various situations, no matter what channel it is. The fundamental idea of this paper is to combine multiband fusion with adaptive contrast stretch (ACS) dehazing approaches, resulting in balanced image improvement while elaborating on visual details. The foundation and detail layers for intensity and Laplacian modules are extracted using multiband decomposition. For the intensity module, the proposed ambient map and transmission estimation can restore the correct intensity. Details on each residual layer are adjusted using adaptive nonlinear mapping algorithms. Our results show excellent performance on various hazy photos using color-corrected reconstruction. Furthermore, we use ACS based on a modified histogram equalization procedure. It is adaptable and gives a localized contrast enhancement effect that is impossible to achieve with typical contrast stretching methods. The proposed method has been thoroughly tested in terms of traditional image quality comparison.