Proceedings of the 20th Conference on Computer Science and Intelligence Systems (FedCSIS)

DOI: http://dx.doi.org/10.15439/2025F5576

Citation: Proceedings of the 20th Conference on Computer Science and Intelligence Systems (FedCSIS), M. Bolanowski, M. Ganzha, L. Maciaszek, M. Paprzycki, D. Ślęzak (eds). ACSIS, Vol. 43, pages 363–368 (2025)

Full text

Abstract. We present a novel multi-stage method for colour image segmentation, with a primary focus on vessels segmentation in retinal fundus images, using a U-Net based architecture. Our approach tackles challenges posed by varying image resolutions through a coarse-to-fine segmentation pipeline. It begins with a rough segmentation at varying scales, guided by a traditional CNN and progressively refines results to find a target resolution. It culminates with detailed segmentation at a target scale with a smaller window sliding step, compared to previous stages. We train and validate our method using four publicly available datasets FIVES, DRHAGIS, HRF, and STARE---and demonstrate superior performance compared to traditional sliding window techniques. Notably, our model achieves high accuracy with relatively few training images. The entire framework is opensourced and adaptable to a wide range of image segmentation tasks.

References

1. Krizhevsky, Alex and Sutskever, Ilya and Hinton, Geoffrey E.,ImageNet Classification with Deep Convolutional Neural Networks,Advances in Neural Information Processing Systems (NIPS), volume 25, pages 1097–1105, 2012,https://dx.doi.org/10.1145/3065386 https://proceedings.neurips.cc/paper/2012/file/c399862d3b9d6b76c8436e924a68c45b-Paper.pdf
2. Ronneberger, Olaf and Fischer, Philipp and Brox, Thomas,U-Net: Convolutional Networks for Biomedical Image Segmentation,Medical Image Computing and Computer-Assisted Intervention (MICCAI),pages 234–241, 2015, Springer,https://dx.doi.org/10.1007/978-3-319-24574-4_28,https://arxiv.org/abs/1505.04597
3. Lesław M. Pawlaczyk, U-Net segmentation implementation,2025, https://github.com/palles77/UnetSegmentation,Accessed: January 29, 2025
4. Ren, Kan and Chang, Longdan and Wan, Minjie and Gu, Guohua and Chen, Qian,An improved U-net based retinal vessel image segmentation method,Heliyon, volume 8, number 10, pages e11187, 2022, Elsevier, https://dx.doi.org/10.1016/j.heliyon.2022.e11187, https://www.sciencedirect.com/science/article/pii/S2405844022024756
5. Liu, Rui and Pu, Wen and Nan, Hong and Zou, Yao,Retina image segmentation using the three-path Unet model,Scientific Reports, volume 13, number 1, pages 22579, 2023, Nature Pub- lishing Group,https://dx.doi.org/10.1038/s41598-023-50141-0, https://www.nature.com/articles/s41598-023-50141-0
6. Das, Smita and Chakraborty, Suvadip and Mishra, Madhusudhan and Majumder, Swanirbhar,Assessment of retinal blood vessel segmentation using U-Net model: A deep learning approach,Franklin Open, volume 8, pages 100143, 2024, Elsevier,https://dx.doi.org/10.1016/j.fraope.2024.100143, https://www.sciencedirect.com/science/article/pii/S2773186324000732
7. Yun, Jiang and Wang, Falin and Gao, Jing and Cao, Simin,Multi-Path Recurrent U-Net Segmentation of Retinal Fundus Image,Applied Sciences, volume 10, number 11, pages 3777, 2020, MDPI, https://dx.doi.org/10.3390/app10113777, https://www.researchgate.net/publication/341810610_Multi-Path_Recurrent_U-Net_Segmentation_of_Retinal_Fundus_Image
8. Huang, Ko-Wei and Yang, Yao-Ren and Huang, Zih-Hao and Liu, Yi-Yang and Lee, Shih-Hsiung,Retinal Vascular Image Segmentation Using Improved UNet Based on Residual Module,Bioengineering, volume 10, number 6, pages 722, 2023, MDPI, https://dx.doi.org/10.3390/bioengineering10060722, https://www.mdpi.com/2306-5354/10/6/722
9. Westwańska, W. and Respondek, J., Counting instances of objects in color images using U-Net network on example of honey bees, in Proceedings of the 2019 Federated Conference on Computer Science and Information Systems (FedCSIS), Annals of Computer Science and Information Systems, vol.18, pp.87–90, Leipzig, Germany, September2019, https://dx.doi.org/10.15439/2019F94, https://ieeexplore.ieee.org/document/8859742
10. Kai Jin, Xingru Huang, Jingxing Zhou, Yunxiang Li, Yan Yan, Yibao Sun, Qianni Zhang, Yaqi Wang, Juan Ye, FIVES: A Fundus Image Dataset for Artificial Intelligence based Vessel Segmentation, 2022 Data Descriptor in Scientific Data, volume 9, number 475, pages 475, 2022, Nature Portfolio, https://dx.doi.org/10.1038/s41597-022-01590-2, https://www.nature.com/articles/s41597-022-01564-3, https://figshare.com/ndownloader/files/34969398
11. Holm, Sven, Russell, Greg, Nourrit, Vincent, McLoughlin, Niall, DR HAGIS - a fundus image database for the automatic extraction of retinal surface vessels from diabetic patients, 2017 Journal of Medical Imaging (Bellingham), volume 4, number 1, pages 014503, 2017, SPIE, https://dx.doi.org/10.1117/1.JMI.4.1.014503, https://pmc.ncbi.nlm.nih.gov/articles/PMC5299858/pdf/JMI-004-014503.pdf, http://personalpages.manchester.ac.uk/staff/niall.p.mcloughlin/DRHAGIS.zip
12. Budai, Attila, Bock, Rüdiger, Maier, Andreas, Hornegger, Joachim, Michelson, Georg,Robust Vessel Segmentation in Fundus Images,2013 International Journal of Biomedical Imaging,volume 2013, article 154860, 2013, Hindawi, https://dx.doi.org/10.1155/2013/154860, http://www5.informatik.uni-erlangen.de/Forschung/Publikationen/2013/Budai13-RVS.pdf, https://www5.cs.fau.de/fileadmin/research/datasets/fundus-images/all.zip
13. Hoover, A. D., Kouznetsova, V., Goldbaum, M.,Locating blood vessels in retinal images by piecewise threshold probing of a matched filter response,2000 IEEE Transactions on Medical Imaging,volume 19, number 3, pages 203–210, 2000, IEEE, https://dx.doi.org/10.1109/42.845178, https://cecas.clemson.edu/~ahoover/stare/probing/index.html
14. Jaccard, P., Étude comparative de la distribution florale dans une portion des Alpes et des Jura, 1901, Bulletin de la Société Vaudoise des Sciences Naturelles, volume 37, pages 547–579, Société Vaudoise des Sciences Naturelles.