HazeNet: a network for single image dehazing

Abstract

In this letter, we present a novel integrated feature that incorporates traditional parameters, and adopt a parallel cascading fashion network HazeNet for enhancing image quality. Our unified feature is a complete integration, and its role is to directly describe the effects of haze. In HazeNet, we design two separate structures including backbone and auxiliary networks to extract feature map. Backbone network is responsible for extracting high-level feature map, and low-level feature learned by the auxiliary network can be interpreted as fine-grained feature. After cascading two features with different accuracy, final performance can be effectively improved. Extensive experimental results on both synthetic datasets and real-world images prove the superiority of the proposed method, and demonstrate more favorable performance compared with the existing state-of-art methods.

References

[1] WangM W, ZhuF Z, BaiY Y. An improved image blind deblurring based on dark channel prior[J]. Optoelectronics letters, 2021, 17(1):40-46

[2] YangY, WangZ W. Haze removal: push DCP at the edge[J]. IEEE signal processing letters, 2020, 27: 1405-1409

[3] GuoF, ZhouC, LiuL J, et al.. Single image defogging based on particle swarm optimization[J]. Optoelectronics letters, 2017, 13(6):452-456

[4] HeK M, SunJ, TangX O. Single image haze removal using dark channel prior[J]. IEEE transactions on pattern analysis and machine intelligence, 2011, 33(12):2341-2353

[5] ZhuQ S, MaiJ M, ShaoL. A fast single image haze removal algorithm using color attenuation prior[J]. IEEE transactions on image processing, 2015, 24(11):3522-3533

[6] MengG F, WangY, DuanJ Y, et al.. Efficient image dehazing with boundary constraint and contextual regularization[C], 2013, New York, IEEE: 617-624

[7] RenW Q, LiuS, ZhangH, et al.. Single image dehazing via multi-scale convolutional neural networks[C], 2016, Berlin, Springer: 154-169

[8] QinX, WangZ L, BaiY C, et al.. FFA-Net: feature fusion attention network for single image dehazing[C], 2020, New York, AAAI: 11908-11915

[9] LiB Y, PengX L, WangZ Y, et al.. AOD-Net: all-in-one dehazing network[C], 2017, New York, IEEE: 4780-4788

[10] WuQ B, ZhangJ G, RenW Q, et al.. Accurate transmission estimation for removing haze and noise from a single image[J]. IEEE transactions on image processing, 2020, 29: 2583-2597

[11] RenW Q, ZhangJ G, XuX Y, et al.. Deep video dehazing with semantic segmentation[J]. IEEE transactions on image processing, 2019, 28(04):1895-1908

[12] LiB Y, RenW Q, FuD P, et al.. Benchmarking single-image dehazing and beyond[J]. IEEE transactions on image processing, 2019, 28(1):492-505

[13] CaiB L, XuX M, JiaK, et al.. DehazeNet: an end-to-end system for single image haze removal[J]. IEEE transactions on image processing, 2016, 25(11):5187-5198

[14] ZhangY F, DingL, SharmaG. Hazerd: an outdoor scene dataset and benchmark for single image dehazing[C], 2017, New York, IEEE: 3205-3209