[1]郭 嘉,蒋 旻,刘双元,等.基于自注意力机制的单幅图像去雨滴方法[J].计算机技术与发展,2021,31(05):54-61.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 010]
 ,,et al.RaindropRemovalMethodforSingleImageBasedonSelf-attentionMechanism[J].,2021,31(05):54-61.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 010]
点击复制

基于自注意力机制的单幅图像去雨滴方法()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
31
期数:
2021年05期
页码:
54-61
栏目:
图形与图像
出版日期:
2021-05-10

文章信息/Info

Title:
RaindropRemovalMethodforSingleImageBasedonSelf-attentionMechanism
文章编号:
1673-629X(2021)05-0054-08
作者:
郭 嘉12蒋 旻12刘双元12江佳俊12
1.武汉科技大学计算机科学与技术学院,湖北武汉430065
2.智能信息处理与实时工业系统湖北省重点实验室(武汉科技大学),湖北武汉430065
Author(s):
GUOJiaJIANGMinLIUShuang-yuanJIANGJia-jun
1.SchoolofComputerScienceandTechnology,WuhanUniversityofScienceandTechnology,Wuhan430065,China;2.HubeiProvinceKeyLaboratoryofIntelligentInformationProcessingandReal-timeIndustrialSystem(WuhanUniversityofScienceandTechnology),Wuhan430065,China
关键词:
雨滴去除深度学习图像去噪轻量级算法自注意力级联稠密残差网络
Keywords:
raindropremovaldeeplearningimagedenoisinglightweightalgorithmself-attentioncascadeddenseresidualnetwork
分类号:
TP391.41
DOI:
10. 3969 / j. issn. 1673-629X. 2021. 05. 010
摘要:
在数字图像中去除雨滴的干扰,对恢复图像质量有较大应用意义。随着深度学习图像去噪技术的发展,针对目前大多数去雨滴的方法恢复质量低、计算量大等问题,提出了一种基于自注意力机制的轻型图像去雨滴算法。该算法建立了一个轻量级的级联稠密残差网络(cascadeddenseresidualnetwork),用于恢复被雨滴覆盖的图像信息。该网络由多个模块组成,模块间用稠密的残差和跳过连接引导图像信息的输出,以从粗略到细节的方式逐级消除图像中的雨滴,恢复雨滴下的背景信息。网络中每个模块由卷积层、非局部神经网络(non-localneuralnetwork)和递归卷积网络组成,在保证预测无雨图像的效果的同时减少参数量。实验结果表明,与AttentiveGAN等去雨滴方法相比,该算法去雨滴效果良好。该方法将自注意力机制加入级联稠密残差网络中,参数量仅为0.22M,适用于小型嵌入式的除雨滴设备。
Abstract:
Removalofraindropinterferenceindigitalimagesisofgreatsignificanceforrestoringimagequality.Withthedevelopmentofdeeplearningimagedenoisingtechnology,inordertosolvetheproblemsoflowrecoveryqualityandlargecalculationamountofmostcurrentraindropremovalmethods,alightimageraindropremovalalgorithm basedonself-attentionmechanism isproposed.Thealgorithmestablishesalightweightcascadeddenseresidualnetworktorecoverimageinformationcoveredbyraindrops.Thenetworkiscomposedofmultiplemodules.Denseresidualsandskipconnectionsareusedbetweenthemodulestoguidetheoutputofimageinformation.Theraindropsintheimageareeliminatedstepbystepfromtheroughtothedetails,andthebackgroundinformationundertheraindropsisrestored.Eachmoduleinthenetworkiscomposedofconvolutionallayers,thenon-localneuralnetworkandtherecursiveconvolutionnetwork,whichreducesthenumberofparameterswhileensuringtheeffectofpredictingde-raindropimages.ExperimentalresultsshowthatcomparedwiththeraindropremovalmethodsuchasAttentiveGAN,thealgorithm hasabetterraindropremovaleffect.Inthismethod,theself-attentionmechanismisaddedtothecascadeddenseresidualnetwork.Theparameteristheonly0.22M,whichissuitableforsmallembeddedraindropdevices.

相似文献/References:

[1]陈强锐,谢世朋.基于深度学习的肺部肿瘤检测方法[J].计算机技术与发展,2018,28(04):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
 CHEN Qiang-rui,XIE Shi-peng.Lung Cancer Detection Method Based on Deep Learning[J].,2018,28(05):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
[2]施泽浩,赵启军.基于全卷积网络的目标检测算法[J].计算机技术与发展,2018,28(05):55.[doi:10.3969/j.issn.1673-629X.2018.05.013]
 SHI Ze-hao,ZHAO Qi-jun.Object Detection Algorithm Based on Fully Convolutional Neural Network[J].,2018,28(05):55.[doi:10.3969/j.issn.1673-629X.2018.05.013]
[3]黄法秀,张世杰,吴志红,等.数据增广下的人脸识别研究[J].计算机技术与发展,2020,30(03):67.[doi:10. 3969 / j. issn. 1673-629X. 2020. 03. 013]
 HUANG Fa-xiu,ZHANG Shi-jie,WU Zhi-hong,et al.Research on Face Recognition Based on Data Augmentation[J].,2020,30(05):67.[doi:10. 3969 / j. issn. 1673-629X. 2020. 03. 013]
[4]陈浩翔,蔡建明,刘铿然,等. 手写数字深度特征学习与识别[J].计算机技术与发展,2016,26(07):19.
 CHEN Hao-xiang,CAI Jian-ming,LIU Keng-ran,et al. Deep Learning and Recognition of Handwritten Numeral Features[J].,2016,26(05):19.
[5]高翔,陈志,岳文静,等.基于视频场景深度学习的人物语义识别模型[J].计算机技术与发展,2018,28(06):53.[doi:10.3969/ j. issn.1673-629X.2018.06.012]
 GAO Xiang,CHEN Zhi,YUE Wen-jing,et al.Human Semantic Recognition Model Based on Video Scene Deep Learning[J].,2018,28(05):53.[doi:10.3969/ j. issn.1673-629X.2018.06.012]
[6]贺飞翔,赵启军. 基于深度学习的头部姿态估计[J].计算机技术与发展,2016,26(11):1.
 HE Fei-xiang,ZHAO Qi-jun. Head Pose Estimation Based on Deep Learning[J].,2016,26(05):1.
[7]徐 融,邱晓晖.一种改进的 YOLO V3 目标检测方法[J].计算机技术与发展,2020,30(07):30.[doi:10. 3969 / j. issn. 1673-629X. 2020. 07. 007]
 XU Rong,QIU Xiao-hui.An Improved YOLO V3 Object Detection[J].,2020,30(05):30.[doi:10. 3969 / j. issn. 1673-629X. 2020. 07. 007]
[8]曾志平[] [],萧海东[],张新鹏[]. 基于DBN的金融时序数据建模与决策[J].计算机技术与发展,2017,27(04):1.
 ZENG Zhi-ping[] [],XIAO Hai-dong[],ZHANG Xin-peng[]. Modeling and Decision-making of Financial Time Series Data with DBN[J].,2017,27(05):1.
[9]李全兵,文 钊*,田艳梅*,等.基于 WGAN 的音频关键词识别研究[J].计算机技术与发展,2021,31(08):26.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 005]
 LI Quan-bing,WEN Zhao *,TIAN Yan-mei *,et al.Research on Audio Keywords Recognition Based on WassersteinGenerative Adversarial Network[J].,2021,31(05):26.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 005]
[10]李宏林. 分析式纹理合成技术及其在深度学习的应用[J].计算机技术与发展,2017,27(11):7.
 LI Hong-lin. Analyzed Texture-synthesis Techniques and Their Applications in Deep Learning[J].,2017,27(05):7.

更新日期/Last Update: 2020-05-10