In this letter, we propose a deep neural network and semi-supervised learning based dehazing algorithm. The dehazing network uses a pyramidal architecture to recover the haze-free scene from a single hazy image in a coarse-to-fine order. To faithfully restore the objects with different scales, we incorporate cascaded multi-scale convolutional blocks into each level of the pyramid. Feature fusion and transfer in the network are achieved using the paths constructed by interleaved residual connections. For better generalization to the complicated haze in real-world environments, we also devise a discriminator that enables semi-supervised adversarial training. Experimental results demonstrate that the proposed work outperforms comparative ones with higher quantitative metrics and more visually pleasant outputs. It can also enhance the robustness of object detection under haze.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Haoyu XU, Yuenan LI, "A Deep Neural Network for Coarse-to-Fine Image Dehazing with Interleaved Residual Connections and Semi-Supervised Training" in IEICE TRANSACTIONS on Information,
vol. E105-D, no. 5, pp. 1125-1129, May 2022, doi: 10.1587/transinf.2021EDL8052.
Abstract: In this letter, we propose a deep neural network and semi-supervised learning based dehazing algorithm. The dehazing network uses a pyramidal architecture to recover the haze-free scene from a single hazy image in a coarse-to-fine order. To faithfully restore the objects with different scales, we incorporate cascaded multi-scale convolutional blocks into each level of the pyramid. Feature fusion and transfer in the network are achieved using the paths constructed by interleaved residual connections. For better generalization to the complicated haze in real-world environments, we also devise a discriminator that enables semi-supervised adversarial training. Experimental results demonstrate that the proposed work outperforms comparative ones with higher quantitative metrics and more visually pleasant outputs. It can also enhance the robustness of object detection under haze.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2021EDL8052/_p
Copy
@ARTICLE{e105-d_5_1125,
author={Haoyu XU, Yuenan LI, },
journal={IEICE TRANSACTIONS on Information},
title={A Deep Neural Network for Coarse-to-Fine Image Dehazing with Interleaved Residual Connections and Semi-Supervised Training},
year={2022},
volume={E105-D},
number={5},
pages={1125-1129},
abstract={In this letter, we propose a deep neural network and semi-supervised learning based dehazing algorithm. The dehazing network uses a pyramidal architecture to recover the haze-free scene from a single hazy image in a coarse-to-fine order. To faithfully restore the objects with different scales, we incorporate cascaded multi-scale convolutional blocks into each level of the pyramid. Feature fusion and transfer in the network are achieved using the paths constructed by interleaved residual connections. For better generalization to the complicated haze in real-world environments, we also devise a discriminator that enables semi-supervised adversarial training. Experimental results demonstrate that the proposed work outperforms comparative ones with higher quantitative metrics and more visually pleasant outputs. It can also enhance the robustness of object detection under haze.},
keywords={},
doi={10.1587/transinf.2021EDL8052},
ISSN={1745-1361},
month={May},}
Copy
TY - JOUR
TI - A Deep Neural Network for Coarse-to-Fine Image Dehazing with Interleaved Residual Connections and Semi-Supervised Training
T2 - IEICE TRANSACTIONS on Information
SP - 1125
EP - 1129
AU - Haoyu XU
AU - Yuenan LI
PY - 2022
DO - 10.1587/transinf.2021EDL8052
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E105-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2022
AB - In this letter, we propose a deep neural network and semi-supervised learning based dehazing algorithm. The dehazing network uses a pyramidal architecture to recover the haze-free scene from a single hazy image in a coarse-to-fine order. To faithfully restore the objects with different scales, we incorporate cascaded multi-scale convolutional blocks into each level of the pyramid. Feature fusion and transfer in the network are achieved using the paths constructed by interleaved residual connections. For better generalization to the complicated haze in real-world environments, we also devise a discriminator that enables semi-supervised adversarial training. Experimental results demonstrate that the proposed work outperforms comparative ones with higher quantitative metrics and more visually pleasant outputs. It can also enhance the robustness of object detection under haze.
ER -