| Title |
Image Dehazing from Depth Estimation using Convolutional Neural Networks |
| Authors |
(Yejin Kim) ; (Changhoon Yim) |
| DOI |
https://doi.org/10.5573/IEIESPC.2021.10.2.075 |
| Keywords |
Image dehazing; Atmospheric scattering model; Deep learning; Depth estimation; Transmission map; Convolutional neural networks |
| Abstract |
Image dehazing is a process that removes haze effects from images. In this paper, we propose a new paradigm that uses depth information to generate the transmission map for image dehazing. Other methods for image dehazing estimate the transmission map directly. We propose the generation of transmission map indirectly from the depth map using the properties of an atmospheric scattering model to obtain a dehazed image. Depth information can be estimated using established methods that employ convolutional neural networks (CNNs) based on deep learning.
Experimental results show that more sophisticated transmission maps can be generated from depth maps without any post-processing. The proposed method can provide improved subjective image dehazing results without any darkening effects compared to previous methods. |