融合特征增强及多尺度损失的增量去雾算法

doi:10.13190/j.jbupt.2021-106

北京邮电大学学报 ›› 2022, Vol. 45 ›› Issue (2): 57-64.doi: 10.13190/j.jbupt.2021-106

融合特征增强及多尺度损失的增量去雾算法

王科平¹, 韦金阳¹, 杨艺^1,2, 费树岷³, 崔科飞²

1. 河南理工大学电气工程与自动化学院, 焦作 454003;
2. 郑州煤矿机械集团股份有限公司郑煤机液压电控有限公司, 郑州 450016;
3. 东南大学自动化学院, 南京 210096

收稿日期:2021-05-28 发布日期:2021-12-16
通讯作者: 韦金阳(1996—),男,硕士生,邮箱:1020763449@qq.com。 E-mail:1020763449@qq.com
作者简介:王科平(1976—),女,副教授,硕士生导师。
基金资助:
国家重点研发计划项目(2018YFC0604502);河南省科技攻关项目(212102210390,192102210100);河南省煤矿智能开采技术创新中心支撑项目(2021YD01);贵州省科技计划资助项目(黔科合重大专项字[2018]3003-1)

Incremental Dehazing Algorithm Combining Feature Enhancement and Multi-Scale Loss

WANG Keping¹, WEI Jinyang¹, YANG Yi^1,2, FEI Shumin³, CUI Kefei²

1. School of Electrical Engineering and Automation, Henan Polytechnic University, Jiaozuo 454003, China;
2. Zhengmeiji Hydraulic Electric Control Company Limited, Zhengzhou Coal Mining Machinery Group Company Limited, Zhengzhou 450016, China;
3. College of Automation, Southeast University, Nanjing 210096, China

Received:2021-05-28 Published:2021-12-16

摘要/Abstract

摘要： 为提高雾霾图像清晰化效果并实现去雾能力的泛化,提出一种特征增强及多尺度损失约束的网络结构,并采用增量式训练方法对网络进行训练。网络由教师网络和学生网络构成。通过学习教师网络提取的标注样本注意力信息对学生网络提取的特征进行特征增强;将标注样本多尺度语义特征作为软标签,建立多尺度语义特征损失衡量机制,与全局像素差异损失级联,构建面向特征和像素的损失函数;采用增量式训练方法,教师网络为学生网络平衡不同数据集的新旧知识提供先验约束,使网络保留原有知识的前提下,快速提高对增补数据集的泛化能力。实验结果表明,所提算法在主观视觉效果与客观评价指标上均取得了较好的效果。

关键词: 图像去雾, 特征增强, 多尺度损失约束, 增量式训练

Abstract: To improve the clarity of the haze image and realize the generalization of the dehazing ability, a network structure with feature enhancement and multi-scale loss constraint is proposed, which is trained by an incremental training method. The network consists of a teacher network and a student network. The student network enhances the features by learning the attention information of the labeled samples extracted by the teacher network and uses the multi-scale semantic features of the labeled samples as the soft targets. Besides, a multi-scale semantic feature loss measurement mechanism is established, which cascades with the global pixel difference loss to construct the loss function for feature and pixel levels. According to the incremental training method, the teacher network guides the student network to balance the relationship between the new and old knowledge of different datasets. Thus, the network can improve the generalization ability of the supplementary dataset quickly, while retains the original knowledge. Experiment results show that the proposed algorithm performs well in both subjective visual effects and objective evaluation indicators.

Key words: image dehazing, feature enhancement, multi-scale loss constraint, incremental training

中图分类号:

TP391.4

王科平, 韦金阳, 杨艺, 费树岷, 崔科飞. 融合特征增强及多尺度损失的增量去雾算法[J]. 北京邮电大学学报, 2022, 45(2): 57-64.

WANG Keping, WEI Jinyang, YANG Yi, FEI Shumin, CUI Kefei. Incremental Dehazing Algorithm Combining Feature Enhancement and Multi-Scale Loss[J]. Journal of Beijing University of Posts and Telecommunications, 2022, 45(2): 57-64.

参考文献

[1] 陈龙,郭宝龙,毕娟,等.基于联合双边滤波的单幅图像去雾算法[J].北京邮电大学学报, 2012, 35(4):19-23. CHEN L, GUO B L, BI J, et al. Algorithm of single image fog removal based on joint bilateral filter[J]. Journal of Beijing University of Posts and Telecommunications, 2012, 35(4):19-23.
[2] 夏璞,刘学斌.大气散射校正静态实时偏振图像去雾技术[J].北京邮电大学学报, 2016, 39(6):33-36, 71. XIA P, LIU X B. Real-time static polarimetric image dehazing technique based on atmospheric scattering correction[J]. Journal of Beijing University of Posts and Telecommunications, 2016, 39(6):33-36, 71.
[3] HE K, JIAN S, TANG X, et al. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12):2341-2353.
[4] CAI B L, XU X M, JIA K, et al. DehazeNet:an end-to-end system for single image haze removal[J]. IEEE Transactions on Image Processing, 2016, 25(11):5187-5198.
[5] REN W Q, LIU S, ZHANG H, et al. Single image dehazing via multi-scale convolutional neural network[J]. European Conference on Computer Vision, 2016:154-169.
[6] LI B Y, PENG X L, WANG Z Y, et al. AOD-net:all-in-one dehazing network[C]//2017 IEEE International Conference on Computer Vision (ICCV). Piscataway, NJ:IEEE Press, 2017:4780-4788.
[7] QIN X, WANG Z L, BAI Y C, et al. FFA-eet:feature fusion attention network for single image dehazing[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(7):11908-11915.
[8] WANG K P, YANG Y, LI B F, et al. Uneven image dehazing by heterogeneous twin network[J]. IEEE Access, 2020, 8:118485-118496.
[9] GIRSHICK R. Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision. Piscataway, NJ:IEEE Press, 2015:1440-1448.
[10] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas:IEEE Computer Society, 2016:770-778.
[11] ZHOU W, BOVIK A C, SHEIKH H R, et al. Image quality assessment:from error visibility to structural similarity[J]. IEEE Transactions on Image Processing:a Publication of the IEEE Signal Processing Society, 2004, 13(4):600-612.
[12] ZEILER M D. ADADELTA:an adaptive learning rate method[EB/OL].[2021-03-20]. https://arxiv.org/abs/1212.5701.
[13] LI B, REN W, FU D, et al. Benchmarking single-image dehazing and beyond[J]. IEEE Transactions on Image Processing:a Publication of the IEEE Signal Processing Society, 2018, 28(1):492-505.
[14] HUYNHTHU Q, GHANBARI M. Scope of validity of PSNR in image/video quality assessment[J]. Electronics Letters, 2008, 44(13):800-801.

融合特征增强及多尺度损失的增量去雾算法

Incremental Dehazing Algorithm Combining Feature Enhancement and Multi-Scale Loss

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 3

编辑推荐

Metrics

本文评价

[1]	刘逸琛, 孙华志, 马春梅, 姜丽芬, 钟长鸿. 一种基于高层特征融合的网络商品分类[J]. 北京邮电大学学报, 2020, 43(5): 98-104,117.
[2]	夏璞, 刘学斌. 大气散射校正静态实时偏振图像去雾技术[J]. 北京邮电大学学报, 2016, 39(6): 33-36,71.
[3]	陈龙郭宝龙毕娟朱娟娟. 基于联合双边滤波的单幅图像去雾算法[J]. 北京邮电大学学报, 2012, 35(4): 19-23.