Joint conditional generative adversarial networks for eyelash artifact removal in ultra-wide-field fundus images

被引:2
|
作者
Zhang, Jiong [1 ,2 ]
Sha, Dengfeng [1 ,3 ]
Ma, Yuhui [1 ]
Zhang, Dan [4 ]
Tan, Tao [5 ]
Xu, Xiayu [6 ,7 ]
Yi, Quanyong [2 ]
Zhao, Yitian [1 ,2 ]
机构
[1] Chinese Acad Sci, Cixi Inst Biomed Engn, Ningbo Inst Mat Technol & Engn, Ningbo, Peoples R China
[2] Wenzhou Med Univ, Affiliated Ningbo Eye Hosp, Ningbo, Peoples R China
[3] Ningbo Univ, Fac Elect Engn & Comp Sci, Ningbo, Peoples R China
[4] Ningbo Univ Technol, Sch Cyber Sci & Engn, Ningbo, Peoples R China
[5] Macao Polytech Univ, Faulty Appl Sci, Macau, Peoples R China
[6] Xi An Jiao Tong Univ, Sch Life Sci & Technol, Key Lab Biomed Informat Engn Minist Educ, Xian, Peoples R China
[7] Xi An Jiao Tong Univ, Zhejiang Res Inst, Hangzhou, Peoples R China
关键词
retina; ultra-wide-field fundus images; artifact removal; conditional GAN; vessel segmentation; SHADOW REMOVAL; OPTOS OPTOMAP; SENSITIVITY;
D O I
10.3389/fcell.2023.1181305
中图分类号
Q2 [细胞生物学];
学科分类号
071009 ; 090102 ;
摘要
Background: Ultra-Wide-Field (UWF) fundus imaging is an essential diagnostic tool for identifying ophthalmologic diseases, as it captures detailed retinal structures within a wider field of view (FOV). However, the presence of eyelashes along the edge of the eyelids can cast shadows and obscure the view of fundus imaging, which hinders reliable interpretation and subsequent screening of fundus diseases. Despite its limitations, there are currently no effective methods or datasets available for removing eyelash artifacts from UWF fundus images. This research aims to develop an effective approach for eyelash artifact removal and thus improve the visual quality of UWF fundus images for accurate analysis and diagnosis.Methods: To address this issue, we first constructed two UWF fundus datasets: the paired synthetic eyelashes (PSE) dataset and the unpaired real eyelashes (uPRE) dataset. Then we proposed a deep learning architecture called Joint Conditional Generative Adversarial Networks (JcGAN) to remove eyelash artifacts from UWF fundus images. JcGAN employs a shared generator with two discriminators for joint learning of both real and synthetic eyelash artifacts. Furthermore, we designed a background refinement module that refines background information and is trained with the generator in an end-to-end manner.Results: Experimental results on both PSE and uPRE datasets demonstrate the superiority of the proposed JcGAN over several state-of-the-art deep learning approaches. Compared with the best existing method, JcGAN improves PSNR and SSIM by 4.82% and 0.23%, respectively. In addition, we also verified that eyelash artifact removal via JcGAN could significantly improve vessel segmentation performance in UWF fundus images. Assessment via vessel segmentation illustrates that the sensitivity, Dice coefficient and area under curve (AUC) of ResU-Net have respectively increased by 3.64%, 1.54%, and 1.43% after eyelash artifact removal using JcGAN.Conclusion: The proposed JcGAN effectively removes eyelash artifacts in UWF images, resulting in improved visibility of retinal vessels. Our method can facilitate better processing and analysis of retinal vessels and has the potential to improve diagnostic outcomes.
引用
收藏
页数:13
相关论文
共 50 条
  • [21] Controllable fundus image generation based on conditional generative adversarial networks with mask guidance
    Xiaoxin Guo
    Xiang Li
    Qifeng Lin
    Guangyu Li
    Xiaoying Hu
    Songtian Che
    Multimedia Tools and Applications, 2024, 83 : 46065 - 46085
  • [22] Electronic Map Visible Watermark Removal with Conditional Generative Adversarial Networks
    Jiang B.
    Xu S.
    Wang J.
    Wang M.
    Journal of Geo-Information Science, 2023, 25 (02) : 288 - 297
  • [23] Controllable fundus image generation based on conditional generative adversarial networks with mask guidance
    Guo, Xiaoxin
    Li, Xiang
    Lin, Qifeng
    Li, Guangyu
    Hu, Xiaoying
    Che, Songtian
    MULTIMEDIA TOOLS AND APPLICATIONS, 2023, 83 (15) : 46065 - 46085
  • [24] Diagnostic Availability of Ultra-Wide-field Fundus Imaging in Korean Patient with Retinal Break
    Ahn, Hyun Min
    Rim, Tyler Hyungtaek
    Chung, Eun Jee
    JOURNAL OF THE KOREAN OPHTHALMOLOGICAL SOCIETY, 2016, 57 (08): : 1254 - 1259
  • [25] Metal artifact correction of industrial CT images based on generative adversarial networks
    Jiang, Shibo
    Sun, Yuewen
    Xu, Shuo
    Wu, Zhifang
    Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University, 2022, 43 (12): : 1766 - 1771
  • [26] Generative adversarial networks for specular highlight removal in endoscopic images
    Funke, Isabel
    Bodenstedt, Sebastian
    Riediger, Carina
    Weitz, Juergen
    Speidel, Stefanie
    MEDICAL IMAGING 2018: IMAGE-GUIDED PROCEDURES, ROBOTIC INTERVENTIONS, AND MODELING, 2018, 10576
  • [27] Wasserstein generative adversarial networks for motion artifact removal in dental CT imaging
    Jiang, Changhui
    Zhang, Qiyang
    Ge, Yongshuai
    Liang, Dong
    Yang, Yongfeng
    Liu, Xin
    Zheng, Hairong
    Hu, Zhanli
    MEDICAL IMAGING 2019: PHYSICS OF MEDICAL IMAGING, 2019, 10948
  • [28] Motion artifact removal in coronary CT angiography based on generative adversarial networks
    Zhang, Lu
    Jiang, Beibei
    Chen, Qiang
    Wang, Lingyun
    Zhao, Keke
    Zhang, Yaping
    Vliegenthart, Rozemarijn
    Xie, Xueqian
    EUROPEAN RADIOLOGY, 2023, 33 (01) : 43 - 53
  • [29] Motion artifact removal in coronary CT angiography based on generative adversarial networks
    Lu Zhang
    Beibei Jiang
    Qiang Chen
    Lingyun Wang
    Keke Zhao
    Yaping Zhang
    Rozemarijn Vliegenthart
    Xueqian Xie
    European Radiology, 2023, 33 : 43 - 53
  • [30] ULTRA-WIDE-FIELD FUNDUS AUTOFLUORESCENCE FINDINGS IN PATIENTS WITH ACUTE ZONAL OCCULT OUTER RETINOPATHY
    Shifera, Amde Selassie
    Pennesi, Mark E.
    Yang, Paul
    Lin, Phoebe
    RETINA-THE JOURNAL OF RETINAL AND VITREOUS DISEASES, 2017, 37 (06): : 1104 - 1119