A GAN-based deep enhancer for quality enhancement of retinal images photographed by a handheld fundus camera
A GAN-based deep enhancer for quality enhancement of retinal images photographed by a handheld fundus camera
- 眼科实践与研究新进展 2022年2卷第3期 页码：100077
- 作者机构：
1. ,Beijing,China
2. ,Hunan,China
3. ,Beijing,China
4. ,Zhengzhou,China
5. ,Beijing,China
- 作者简介：
- 基金信息：
DOI：10.1016/j.aopr.2022.100077
中图分类号：
扫 描 看 全 文
A GAN-based deep enhancer for quality enhancement of retinal images photographed by a handheld fundus camera[J]. 眼科实践与研究新进展, 2022,2(3):100077.
Junxia Fu, Lvchen Cao, Shihui Wei, et al. A GAN-based deep enhancer for quality enhancement of retinal images photographed by a handheld fundus camera[J]. AOPR, 2022,2(3):100077.
A GAN-based deep enhancer for quality enhancement of retinal images photographed by a handheld fundus camera[J]. 眼科实践与研究新进展, 2022,2(3):100077. DOI： 10.1016/j.aopr.2022.100077.
Junxia Fu, Lvchen Cao, Shihui Wei, et al. A GAN-based deep enhancer for quality enhancement of retinal images photographed by a handheld fundus camera[J]. AOPR, 2022,2(3):100077. DOI： 10.1016/j.aopr.2022.100077.
摘要
Objective,Due to limited imaging conditions, the quality of fundus images is often unsatisfactory, especially for images photographed by handheld fundus cameras. Here, we have developed an automated method based on combining two mirror-symmetric generative adversarial networks (GANs) for image enhancement.,Methods,A total of 1047 retinal images were included. The raw images were enhanced by a GAN-based deep enhancer and another methods based on luminosity and contrast adjustment. All raw images and enhanced images were anonymously assessed and classified into 6 levels of quality classification by three experienced ophthalmologists. The quality classification and quality change of images were compared. In addition, image-detailed reading results for the number of dubiously pathological fundi were also compared.,Results,After GAN enhancement, 42.9% of images increased their quality, 37.5% remained stable, and 19.6% decreased. After excluding the images at the highest level (level 0) before enhancement, a large number (75.6%) of images showed an increase in quality classification, and only a minority (9.3%) showed a decrease. The GAN-enhanced method was superior for quality improvement over a luminosity and contrast adjustment method (,P,＜0.001). In terms of image reading results, the consistency rate fluctuated from 86.6% to 95.6%, and for the specific disease subtypes, both discrepancy number and discrepancy rate were less than 15 and 15%, for two ophthalmologists.,Conclusions,Learning the style of high-quality retinal images based on the proposed deep enhancer may be an effective way to improve the quality of retinal images photographed by handheld fundus cameras.
Abstract
Objective,Due to limited imaging conditions, the quality of fundus images is often unsatisfactory, especially for images photographed by handheld fundus cameras. Here, we have developed an automated method based on combining two mirror-symmetric generative adversarial networks (GANs) for image enhancement.,Methods,A total of 1047 retinal images were included. The raw images were enhanced by a GAN-based deep enhancer and another methods based on luminosity and contrast adjustment. All raw images and enhanced images were anonymously assessed and classified into 6 levels of quality classification by three experienced ophthalmologists. The quality classification and quality change of images were compared. In addition, image-detailed reading results for the number of dubiously pathological fundi were also compared.,Results,After GAN enhancement, 42.9% of images increased their quality, 37.5% remained stable, and 19.6% decreased. After excluding the images at the highest level (level 0) before enhancement, a large number (75.6%) of images showed an increase in quality classification, and only a minority (9.3%) showed a decrease. The GAN-enhanced method was superior for quality improvement over a luminosity and contrast adjustment method (,P,＜0.001). In terms of image reading results, the consistency rate fluctuated from 86.6% to 95.6%, and for the specific disease subtypes, both discrepancy number and discrepancy rate were less than 15 and 15%, for two ophthalmologists.,Conclusions,Learning the style of high-quality retinal images based on the proposed deep enhancer may be an effective way to improve the quality of retinal images photographed by handheld fundus cameras.
关键词
Keywords
Generative adversarial networks (GANs)Retinal imageQualityHandheld fundus cameras
references
1 M.D. Abràmoff, M.K. Garvin, M. SonkaRetinal imaging and image analysis IEEE Rev Biomed Eng, 3 (2010), pp. 169-208, 10.1109/rbme.2010.2084567
2 Y. Kandasamy, R. Smith, I. Wright, et al.Use of digital retinal imaging in screening for retinopathy of prematurity J Paediatr Child Health, 49 (2013), pp. E1-E5, 10.1111/j.1440-1754.2012.02557.x
3 S. Kankanahalli, P.M. Burlina, Y. Wolfson, et al.Automated classification of severity of age-related macular degeneration from fundus photographs Invest Ophthalmol Vis Sci, 54 (2013), pp. 1789-1796, 10.1167/iovs.12-10928
4 A.S. Coyner, J. Chen, J.P. Campbell, et al.Diagnosability of synthetic retinal fundus images for plus disease detection in retinopathy of prematurity AMIA Annu Symp Proc, undefined (2020), pp. 329-337
5 A. Diaz-Pinto, A. Colomer, V. Naranjo, et al.Retinal image synthesis and semi-supervised learning for glaucoma assessment IEEE Trans Med Imag, 38 (2019), pp. 2211-2218, 10.1109/tmi.2019.2903434
6 W. Wang, W. Yan, A. Müller, et al.Association of socioeconomics with prevalence of visual impairment and blindness JAMA Ophthalmol, 135 (2017), pp. 1295-1302, 10.1001/jamaophthalmol.2017.3449
7 U. Şevik, C. Köse, T. Berber, et al.Identification of suitable fundus images using automated quality assessment methods J Biomed Opt, 19 (2014), Article 046006, 10.1117/1.Jbo.19.4.046006
8 M.R. Mookiah, U.R. Acharya, C.K. Chua, et al.Computer-aided diagnosis of diabetic retinopathy: a review Comput Biol Med, 43 (2013), pp. 2136-2155, 10.1016/j.compbiomed.2013.10.007
9 J.S. Lim, M. Hong, W.S.T. Lam, et al.Novel technical and privacy-preserving technology for artificial intelligence in ophthalmology Curr Opin Ophthalmol (2022), 10.1097/icu.0000000000000846
10 Q.T.M. Pham, S. Ahn, J. Shin, et al.Generating future fundus images for early age-related macular degeneration based on generative adversarial networks Comput Methods Progr Biomed, 216 (2022), Article 106648, 10.1016/j.cmpb.2022.106648
11 C. Zheng, V. Koh, F. Bian, et al.Semi-supervised generative adversarial networks for closed-angle detection on anterior segment optical coherence tomography images: an empirical study with a small training dataset Ann Transl Med, 9 (2021), p. 1073, 10.21037/atm-20-7436
12 C. Zheng, F. Bian, L. Li, et al.Assessment of generative adversarial networks for synthetic anterior segment optical coherence tomography images in closed-angle detection Transl Vis Sci Technol, 10 (2021), p. 34, 10.1167/tvst.10.4.34
13 E. Yildiz, A.T. Arslan, A. Yildiz Tas, et al.Generative adversarial network based automatic segmentation of corneal subbasal nerves on in vivo confocal microscopy images Transl Vis Sci Technol, 10 (2021), p. 33, 10.1167/tvst.10.6.33
14 C. Zheng, X. Xie, K. Zhou, et al.Assessment of generative adversarial networks model for synthetic optical coherence tomography images of retinal disorders Transl Vis Sci Technol, 9 (2020), p. 29, 10.1167/tvst.9.2.29
15 C. Zheng, H. Ye, J. Yang, et al.Development and clinical validation of semi-supervised generative adversarial networks for detection of retinal disorders in optical coherence tomography images using small dataset Asia Pac J Ophthalmol (Phila) (2022), 10.1097/apo.0000000000000498
16 M. Zhou, K. Jin, S. Wang, et al.Color retinal image enhancement based on luminosity and contrast adjustment IEEE Trans Biomed Eng, 65 (2018), pp. 521-527, 10.1109/tbme.2017.2700627
17 L. Xiong, H. Li, L. XuAn approach to evaluate blurriness in retinal images with vitreous opacity for cataract diagnosis J Healthc Eng (2017), Article 5645498, 10.1155/2017/5645498
18 J. Li, L. Xu Romate Ophthalmology (2017), p. 96
19 H. Zhao, B. Yang, L. Cao, H. Li (Eds.), Data-Driven Enhancement of Blurry Retinal Images via Generative Adversarial Networks, Springer, Cham (2019)
20 L. Miao, Y.Q. Zhao, X.H. Wang, et al.Retinal Vessel Enhancement Based on Multi-Scale Top-Hat Transformation and Histogram Fitting Stretching Opt Laser Technol (2014), pp. 56-62
21 S. Chaudhuri, S. Chatterjee, N. Katz, et al.Detection of blood vessels in retinal images using two-dimensional matched filters IEEE Trans Med Imaging (1989), pp. 263-269, 10.1109/42.34715
22 R. Su, C. Sun, C. Zhang, et al.A new method for linear feature and junction enhancement in 2D images based on morphological operation Pattern Recogniti (2014), pp. 3193-3208
23 T. Lin, M. Du, J. XuThe preprocessing of subtraction and the enhancement for biomedical image of retinal blood vessels J Biomed Phys Eng (2003), p. 56, 10.3321/j.issn:1001-5515.2003.01.016
24 L. Xiong, H. Li, L. XuAn enhancement method for color retinal images based on image formation model Comput Methods Programs Biomed, 143 (2017), pp. 137-150, 10.1016/j.cmpb.2017.02.026
25 J. Zhang, H. Li, Q. Nie, et al.A retinal vessel boundary tracking method based on Bayesian theory and multi-scale line detection Comput Med Imaging Graph, 38 (2014), pp. 517-525, 10.1016/j.compmedimag.2014.05.010
0
浏览量
0
下载量
0
CSCD
- Meta-XML下载
- BibTeX下载
- Endnote下载
- RefMan 下载
- NoteExpress下载
- NoteFirst下载
关联资源
相关文章
相关作者
相关机构
- 技术支持由北京北大方正电子有限公司提供 京ICP备09064830号-19
京公网安备11010802024621 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies