计算机科学 ›› 2021, Vol. 48 ›› Issue (6): 145-152.doi: 10.11896/jsjkx.200900109
石恒, 张玲
SHI Heng, ZHANG Ling
摘要: 虽然现有基于深度学习的图像阴影消除方法已取得了一定的进步,但是这些方法主要关注图像本身,没有很好地探索其他额外与阴影相关的信息,因此这些方法常常存在图像纹理模糊、内容不协调等问题。针对这些问题,文中基于生成对抗网络(Generative Adversarial Network,GAN),提出了一种新的阴影消除网络模型。该方法首先从全局上生成一个粗糙的阴影消除结果,再利用与阴影相关的残差信息对粗糙的结果在颜色和细节上进行局部优化,从而获得更加真实自然的无阴影图像。生成网络包含3个编码-解码结构,首先利用第1个编码-解码结构对阴影图像进行整体光照恢复,生成一个初始的阴影消除结果;同时将与阴影相关的残差信息作为辅助信息输入第2个编码-解码器,对初始结果进行进一步优化;为了避免阴影区域出现纹理不协调等问题,最后利用第3个编码-解码器对阴影区域细节纹理进行修正,使得生成的阴影消除图像更加真实自然。对抗网络由Patch鉴别器构成,用来鉴别图像阴影消除结果的真实性。实验结果表明,与目前的图像阴影消除方法相比,无论在阴影区域还是在非阴影区域上所提方法都达到了最佳的RMSE值,且该方法生成的阴影消除图像与真实无阴影图像更加接近,有效解决了纹理模糊等问题,证实了该方法的有效性和可行性。
中图分类号:
[1]KHAN S H,BENNAMOUN M,SOHEL F,et al.AutomaticShadow Detection and Removal from a Single Image[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2016,38(3):431-446. [2]WEI J,LONG C,ZOU H,et al.Shadow Inpainting and Removal Using Generative Adversarial Networks with Slice Convolutions[J].Computer Graphics Forum,2019,38(7):381-392. [3]IIZUKA S,SIMOSERRA E,ISHIKAWA H,et al.Globally and locally consistent image completion[J].ACM Transactions on Graphics,2017,36(4). [4]WANG T,LIU M,ZHU J,et al.High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs[C]//Computer Vision and Pattern Recognition.2018:8798-8807. [5]RAJ N B,VENKATESWARAN N.Single Image Haze Removal using a Generative Adversarial Network.[J].arXiv:1810.09479. [6]ENGIN D,GENC A,EKENEL H K,et al.Cycle-Dehaze:En-hanced CycleGAN for Single Image Dehazing[C]//Computer Vision and Pattern Recognition.2018:825-833. [7]LE H,VICENTE T F,NGUYEN V,et al.A+D Net:Training a Shadow Detector with Adversarial Shadow Attenuation[C]//European Conference on Computer Vision.2018:680-696. [8]QIAN R,TAN R T,YANG W,et al.Attentive Generative Adversarial Network for Raindrop Removal from A Single Image[C]//Computer Vision and Pattern Recognition.2018:2482-2491. [9]ISOLA P,ZHU J,ZHOU T,et al.Image-to-Image Translation with Conditional Adversarial Networks[C]//Computer Vision and Pattern Recognition,2017:5967-5976. [10]LIU Y,LI Y S,ZHANG D P.Shadow elimination method based on chromaticity distortion and texture features[J].Computer Science,2005,32(9):211-214. [11]CHEN R,LI P,HUANG Y,et al.Motion shadow removal algorithm based on multi-feature fusion[J].Computer Science,2018,45(6):291-295. [12]GUO R,DAI Q,HOIEM D,et al.Single-image shadow detection and removal using paired regions[C]//Computer Vision and Pattern Recognition,2011:2033-2040. [13]XIAO C,XIAO D,ZHANG L,et al.Efficient Shadow Removal Using Subregion Matching Illumination Transfer[J].Computer Graphics Forum,2013,32(7):421-430. [14]FINLAYSON G D,HORDLEY S D,LU C,et al.On the remo-val of shadows from images[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2006,28(1):59-68. [15]QU L,TIAN J,HE S,et al.DeshadowNet:A Multi-context Embedding Deep Network for Shadow Removal[C]//Computer Vision and Pattern Recognition.2017:2308-2316. [16]HU X,FU C,ZHU L,et al.Direction-aware Spatial ContextFeatures for Shadow Detection and Removal[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2018:7454-7462. [17]WANG J,LI X,YANG J,et al.Stacked Conditional Generative Adversarial Networks for Jointly Learning Shadow Detection and Shadow Removal[C]//Computer Vision and Pattern Recognition.2018:1788-1797. [18]SIDOROV O.Conditional GANs for Multi-Illuminant ColorConstancy:Revolution or Yet Another Approach? [C]//The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops.IEEE,2019. [19]DING B,LONG C,ZHANG L,et al.ARGAN:Attentive Recurrent Generative Adversarial Network for Shadow Detection and Removal[J].arXiv:1908.01323. [20]ZHANG L,ZHANG Q,XIAO C,et al.Shadow Remover:Image Shadow Removal Based on Illumination Recovering Optimization[J].IEEE Transactions on Image Processing,2015,24(11):4623-4636. [21]ZHANG L,LONG C,ZHANG X,et al.RIS-GAN:Explore Re-sidual and Illumination with Generative Adversarial Networks for Shadow Removal[C]//National Conference on Artificial Intelligence.2020. [22]LIU F Q,LI Z H.Shadow removal method based on light-independent image[J].Journal of Image and Graphics,2007,12(10):1837-1840. [23]XIAO M,HAN C Z.Edge-based motion shadow removal algorithm in indoor video[J].Pattern Recognition and Artificial Intelligence,2006,19(5):640-644. [24]ZHANG H,PATEL V M.Density-Aware Single Image De-raining Using a Multi-stream Dense Network[C]//Computer Vision and Pattern Recognition.2018:695-704. [25]GUO L S,GUO L,JIAO R H,et al.A target detection algorithm based on moving shadows[J].Pattern Recognition and Artificial Intelligence,2007,20(2):180-184. |
[1] | 张佳, 董守斌. 基于评论方面级用户偏好迁移的跨领域推荐算法 Cross-domain Recommendation Based on Review Aspect-level User Preference Transfer 计算机科学, 2022, 49(9): 41-47. https://doi.org/10.11896/jsjkx.220200131 |
[2] | 孙奇, 吉根林, 张杰. 基于非局部注意力生成对抗网络的视频异常事件检测方法 Non-local Attention Based Generative Adversarial Network for Video Abnormal Event Detection 计算机科学, 2022, 49(8): 172-177. https://doi.org/10.11896/jsjkx.210600061 |
[3] | 戴朝霞, 李锦欣, 张向东, 徐旭, 梅林, 张亮. 基于DNGAN的磁共振图像超分辨率重建算法 Super-resolution Reconstruction of MRI Based on DNGAN 计算机科学, 2022, 49(7): 113-119. https://doi.org/10.11896/jsjkx.210600105 |
[4] | 尹文兵, 高戈, 曾邦, 王霄, 陈怡. 基于时频域生成对抗网络的语音增强算法 Speech Enhancement Based on Time-Frequency Domain GAN 计算机科学, 2022, 49(6): 187-192. https://doi.org/10.11896/jsjkx.210500114 |
[5] | 徐辉, 康金梦, 张加万. 基于特征感知的数字壁画复原方法 Digital Mural Inpainting Method Based on Feature Perception 计算机科学, 2022, 49(6): 217-223. https://doi.org/10.11896/jsjkx.210500105 |
[6] | 高志宇, 王天荆, 汪悦, 沈航, 白光伟. 基于生成对抗网络的5G网络流量预测方法 Traffic Prediction Method for 5G Network Based on Generative Adversarial Network 计算机科学, 2022, 49(4): 321-328. https://doi.org/10.11896/jsjkx.210300240 |
[7] | 黎思泉, 万永菁, 蒋翠玲. 基于生成对抗网络去影像的多基频估计算法 Multiple Fundamental Frequency Estimation Algorithm Based on Generative Adversarial Networks for Image Removal 计算机科学, 2022, 49(3): 179-184. https://doi.org/10.11896/jsjkx.201200081 |
[8] | 瞿中, 陈雯. 基于空洞卷积和多特征融合的混凝土路面裂缝检测 Concrete Pavement Crack Detection Based on Dilated Convolution and Multi-features Fusion 计算机科学, 2022, 49(3): 192-196. https://doi.org/10.11896/jsjkx.210100164 |
[9] | 石达, 芦天亮, 杜彦辉, 张建岭, 暴雨轩. 基于改进CycleGAN的人脸性别伪造图像生成模型 Generation Model of Gender-forged Face Image Based on Improved CycleGAN 计算机科学, 2022, 49(2): 31-39. https://doi.org/10.11896/jsjkx.210600012 |
[10] | 唐雨潇, 王斌君. 基于深度生成模型的人脸编辑研究进展 Research Progress of Face Editing Based on Deep Generative Model 计算机科学, 2022, 49(2): 51-61. https://doi.org/10.11896/jsjkx.210400108 |
[11] | 李建, 郭延明, 于天元, 武与伦, 王翔汉, 老松杨. 基于生成对抗网络的多目标类别对抗样本生成算法 Multi-target Category Adversarial Example Generating Algorithm Based on GAN 计算机科学, 2022, 49(2): 83-91. https://doi.org/10.11896/jsjkx.210800130 |
[12] | 谈馨悦, 何小海, 王正勇, 罗晓东, 卿粼波. 基于Transformer交叉注意力的文本生成图像技术 Text-to-Image Generation Technology Based on Transformer Cross Attention 计算机科学, 2022, 49(2): 107-115. https://doi.org/10.11896/jsjkx.210600085 |
[13] | 陈贵强, 何军. 自然场景下遥感图像超分辨率重建算法研究 Study on Super-resolution Reconstruction Algorithm of Remote Sensing Images in Natural Scene 计算机科学, 2022, 49(2): 116-122. https://doi.org/10.11896/jsjkx.210700095 |
[14] | 张玮琪, 汤轶丰, 李林燕, 胡伏原. 基于场景图的段落生成序列图像方法 Image Stream From Paragraph Method Based on Scene Graph 计算机科学, 2022, 49(1): 233-240. https://doi.org/10.11896/jsjkx.201100207 |
[15] | 蒋宗礼, 樊珂, 张津丽. 基于生成对抗网络和元路径的异质网络表示学习 Generative Adversarial Network and Meta-path Based Heterogeneous Network Representation Learning 计算机科学, 2022, 49(1): 133-139. https://doi.org/10.11896/jsjkx.201000179 |
|