Low-light enhancement based on an improved simplified Retinex model via fast illumination map refinement
- PDF / 3,677,159 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 81 Downloads / 165 Views
SHORT PAPER
Low‑light enhancement based on an improved simplified Retinex model via fast illumination map refinement Shijie Hao1 · Xu Han1 · Youming Zhang2 · Lei Xu3 Received: 26 May 2019 / Accepted: 5 September 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Low-light enhancement is an important post-image-processing technique, as it helps to reveal hidden details from dark image regions. In this paper, we propose a fast low-light enhancement model, which is robust to various lighting conditions and imaging noise, and is computationally efficient. By using a fusion-based simplified Retinex model, our model caters to different lighting conditions. In the model, we propose an edge-preserving filter to efficiently refine the estimated illumination map. We also extend our model by equipping it with a very simple denoising step, which effectively prevents the overboosting of imaging noise in the dark regions. We conduct the experiments on public available images as well as the ones collected by ourselves. Visual and quantitative results validate the effectiveness of our model. Keywords Image contrast enhancement · Illumination map refinement · Simplified retinex model
1 Introduction We have witnessed a rapid development of photographing devices, such as DSLR (Digital Single Lens Reflex) camera or intelligent mobile phone. For example, we can obtain an image with fairly good resolution only with a tiny mobile phone. However, it is still very challenging for common displaying devices to represent an illumination range as wide as that of the real world. As shown in Fig. 1, our photographs at hand can have bad image contrast and unclear image details, when they are taken at imperfect illumination conditions,
* Lei Xu [email protected] Shijie Hao [email protected] Xu Han [email protected] Youming Zhang [email protected] 1
Hefei University of Technology, Hefei 230009, Anhui Province, China
2
Northeastern University At Qinhuangdao, Qinhuangdao 066004, Hebei Province, China
3
Shanghai Polytechnic University, Shanghai City 201209, China
e.g., strong light source (backlight) and weak light source (low light). In this context, it is important to improve the contrast of an image with an imperfect illumination condition. Specifically, at first, we need to enhance the visibility of the dark regions on the one hand, and keep the visual naturalness of the originally bright regions on the other hand. Second, considering that large imaging noise often hides in the dark regions, and they tend to be over-boosted along with image details during the enhancement process. So the enhancement model should be robust to the imaging noise. Third, the overall enhancement model is expected to be computationally effective. In this paper, we propose a simple but effective low-light enhancement model to satisfy the above demands. As for the first demand, we adopt a fusion strategy that adaptively weighs between the intermediate enhanced image and the original image. We note that the fusion proce
Data Loading...
