THE MINIMAX APPROACH APPLICATION FOR SEGMENTATION OF RETINAL VESSELS

V. Martsenyuk, N. Milian, R. Milian

Abstract


The paper presents the results of work neural network for segmentation of the fundus vessels using the Tensorflow machine learning library. Training and testing takes place on the public DRIVE data set. The results of work model, namely the recognized blood vessels are presented.

When considering binary cross-entropy as an indicator of efficiency, which is demonstrated in the article, it was determined that the U-Net model with 8x8 tiles is a solution to the problem of minimax ML. In the first step, the value of the loss function is compared for all considered models. In the second step, it is determined that the value of binary cross-entropy for the U-Net model with 8x8 tiles will be the minimum among the maximum characteristics.

When considering training time as an indicator of efficiency, as shown in the relevant table, the U-Net model with 25x25 tiles is a solution to the minimax ML problem. In the first step, we first compare the values of the training time of all the models under consideration. In the second step, it is determined that the time value for the U-Net model with 25x25 tiles will be the minimum among the maximum values.


Full Text:

PDF

References


C. L Srinidhi, P. Aparna, and J. Rajan, ‘‘Recent advancements in retinal vessel segmentation,’’ J. Med. Syst., vol. 41, no. 4, Mar. 2017.

C. Wang, Z. Zhao, Q. Ren, Y. Xu, and Y. Yu, ‘‘Dense U-net based on patchbased learning for retinal vessel segmentation,’’ Entropy, vol. 21, no. 2, p. 168, Feb. 2019.

C. Y.-L. Cheung, M. K. Ikram, C. Chen, and T. Y. Wong, ‘‘Imaging retina to study dementia and stroke,’’ Prog. Retinal Eye Res., vol. 57, pp. 89–107, Mar. 2017.

E. Tuba, L. Mrkela, and M. Tuba, ‘‘Retinal blood vessel segmentation by support vector machine classification,’’ present at the 27th Int. Conf. Radioelektronika (RADIOELEKTRONIKA), 2017. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/7936649.

F. Uslu and A. A. Bharath, ‘‘A recursive Bayesian approach to describe retinal vasculature geometry,’’ Pattern Recognit., vol. 87, pp. 157–169, Mar. 2019.

H. Akil, A. S. Huang, B. A. Francis, S. R. Sadda, and V. Chopra, ‘‘Retinal vessel density from optical coherence tomography angiography to differentiate early glaucoma, pre-perimetric glaucoma and normal eyes,’’ PLoS ONE, vol. 12, no. 2, Feb. 2017, Art. no. e0170476.

H. Lee, M. Lee, H. Chung, and H. C. Kim, ‘‘Quantification of retinal vessel tortuosity in diabetic retinopathy using optical coherence tomography angiography,’’ Retina, vol. 38, no. 5, pp. 976–985, May 2018.

O. Ronneberger, P. Fischer, and T. Brox, ‘‘U-net: Convolutional networks for biomedical image segmentation,’’ presented at the Med. Image Comput. Comput.-Assist. Intervent. (MICCAI), 2015. [Online]. Available: https://link.springer.com/chapter/10.1007/978-3-319-24574-4_28

Q. Jin, Z. Meng, T. D. Pham, Q. Chen, L. Wei, and R. Su, ‘‘DUNet: A deformable network for retinal vessel segmentation,’’ Knowl.-Based Syst., vol. 178, pp. 149–162, Aug. 2019.

S. Feng, Z. Zhuo, and D. Pan, ‘‘CcNet: A cross-connected convolutional network for segmenting retinal vessels using multi-scale features,’’ Neurocomputing, vol. 392, pp. 268–2767, Jun. 2020.

S. Fraser-Bell, R. Symes, and A. Vaze, ‘‘Hypertensive eye disease: A review,’’ Clin. Experim. Ophthalmol., vol. 45, no. 1, pp. 45–53, Jan. 2017.

S. Guo, K. Wang, H. Kang, Y. Zhang, Y. Gao, and T. Li, ‘‘BTSDSN: Deeply supervised neural network with short connections for retinal vessel segmentation,’’ Int. J. Med. Informat., vol. 126, pp. 105–113, Jun. 2019.

X. Gao, Y. Cai, and C. Qiu, ‘‘Retinal blood vessel segmentation based on the Gaussian matched filter and U-net,’’ presented at the CISP-BMEI, 2017. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/8302199.

X. Xiao, S. Lian, and Z. Luo, ‘‘Weighted Res-UNet for highquality retina vessel segmentation,’’ presented at the 9th Int. Conf. Inf. Technol. Med. Educ. (ITME), 2018. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/8589312.


Refbacks

  • There are currently no refbacks.