摘要
Automatic optic disk (OD) segmentation is an important tool for early detection of eye diseases. In this article, we proposed a Res-UNet network by applying residual learning module and other improvements in U-Net for optic disk segmentation in retinal image. Since training data available is insufficient, we enlarge the data set by generating data pieces. Res-UNet is then trained to classify each pixel of the input retinal image. Finally, the predicted probability map is further post-processed with morphological technique to get final OD segmentation result. Experiments on the public DRISHTI-GS data set including comparison with the best known methods show that the proposed model outperforms most existing methods on several metrics.
延伸閱讀
- Chen, H. T., Wang, C. M., Chan, Y. K., Mao, S. F. Y., Chen, Y. F., & Lin, S. F. (2013). Statistics-Based Initial Contour Detection of Optic Disc on a Retinal Fundus Image Using Active Contour Model. Journal of Medical and Biological Engineering, 33(4), 388-393. https://doi.org/10.5405/jmbe.1243
- Dhanalakshmi, S., & Ravichandran, T. (2012). Blood Vessel Segmentation for Retinal Images Based on Am-fm Method. Research Journal of Applied Sciences, Engineering and Technology, 4(24), 5519-5524. https://www.airitilibrary.com/Article/Detail?DocID=20407467-201212-201512090022-201512090022-5519-5524
- Chen, S. (2020). Multi-path Retinal Vessel Segmentation Algorithm Based on Improved U-shaped Network. International Core Journal of Engineering, 6(6), 86-93. https://doi.org/10.6919/ICJE.202006_6(6).0016
- 黃曼菁、方憲瑞（1995）。Optic Disc Pit Associated Nonrhegmatogenous Retinal Detachment。中華民國眼科醫學會雜誌，34(3)，417-422。https://doi.org/10.30048/ACTASOS.199509.0018
- 陳俊男、李鳳利、林嘉理（1994）。Retinitis Pigmentosa Associated with Unilateral Coats' Disease and Optic Disc Drusen-Case Report。中華民國眼科醫學會雜誌，33(2)，246-253。https://doi.org/10.30048/ACTASOS.199406.0017