Abstract
Diabetes is one of the most dangerous illnesses. Sometimes it can be observed faster by experienced ophthalmologist rather than general practitioner. It is related to the fact that it leads to the small changes in the form of exudates visible in retina color images. Such amendments are known under name of “diabetic retinopathy”. However, it needs to be claimed that, in most of the cases, when detected it is too advanced to preserve patient’s eyesight. This is the main reason why it is urgent to work-out novel ideas and approaches that can detect these changes in their early stages. In this work we propose a neural network-based algorithm for diabetic retinopathy recognition. For the aim of classification, we consumed well-known CNN’s architectures as ResNet50 or VGG-16 but also, new ones as InceptionV4. The best results reached 95% in the task of diabetic retinopathy recognition. Moreover, to increase confidence, saliency maps were introduced – by this solution we observed which parts of the image had the highest impact on the classifier decisions. Our observations confirmed that exudates were the most important in the classification process.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
https://openai.com/blog/chatgpt. Accessed 11 Feb 2023
Quer, G., Arnaout, R., Henne, M., Arnaout, R.: Machine learning and the future of cardiovascular care: JACC state-of-the-art review. J. Am. Coll. Cardiol. 77(3), 300–313 (2021)
Chung, J., Teo, J.: Mental health prediction using machine learning: taxonomy, applications and challenges. Appl. Comput. Intell. Soft Comput. 2022, Article no. 9970363 (2022). https://doi.org/10.1155/2022/9970363
Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. MIT Press, Cambridge (2016)
Molnar, C.: Interpretable machine learning: a guide for making black box models explainable (2022). ISBN: 979-8411463330
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. arXiv: 1512.03385 [cs.CV], 10 December 2015
Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. arXiv: 1409.1556 [cs.CV], 10 April 2015
Szegedy, C., Ioffe, S., Vanhoucke, V., Alemi, A.: Inception-v4, Inception-ResNet and the impact of residual connections on learning. arXiv: 1602.07261v2 [cs.CV], 23 August 2016
Monemian, M., Rabbani, H.: Exudate Identification in retinal fundus images using precise textural verifications. Sci. Rep. 13, Article no. 2824 (2023)
Malhi, A., Grewal, R., Pannu H.: Detection and diabetic retinopathy grading using digital retinal images. Int. J. Intell. Robot. Appl. (2023)
Basit, A., Moazam Fraz, M.: Optic disk detection and boundary extraction in retinal images. Appl. Opt. 54(11), 3440–3447 (2015)
Bahadar, K., Khaliq, A., Shahid, M.: A morphological hessian based approach for retinal blood vessels segmentation and denoising using region based otsu thresholding. PLoS ONE 11(7) (2016). PMCID: PMC4956315
Galdran, A., Anjos, A., Dolz, J., Chakor, H., Lombaert, H., Ayed, I.: State-of-the-art retinal vessel segmentation with minimalistic models. Sci. Rep. 12, Article no. 6174 (2022)
https://www.projectpro.io/article/anomaly-detection-using-machine-learning-in-python-with-example/555. Accessed 11 Feb 2023
Du, Y.: Anomaly detection in fundus images by self-adaptive decomposition via local and color based sparse coding. Biomed. Opt. Express 13(8), 4261–4277 (2022)
Gadekallu, T., Khare, N., Bhattacharya, S., et al.: Deep neural networks to predict diabetic retinopathy. J. Ambient Intell. Humaniz. Comput. (2020). https://doi.org/10.1007/s12652-020-01963-7
Wu, H., Zhao, S., Zhang, X., et al.: Back-propagation artificial neural network for early diabetic retinopathy detection based on a priori knowledge. J. Phys. Conf. Ser. Article no. 012019 (2020). https://doi.org/10.1088/1742-6596/1437/1/012019
Benzamin, A., Chakraborty, C.: Detection of hard exudates in retinal fundus images using deep learning. arXiv: 1808.03656 [cs.CV] (2018)
Patil, A., Chakravorty, C.: Detection of hard exudate using retinal optical coherence tomography (OCT) images. Glob. Transit. Proc. 2(2), 566–570 (2021)
Ekman, M.: Learning Deep Learning. Addison-Wesley (2022). ISBN: 978-0-13-747035-8
Yasashvini, R., Vergin Raja Sarobin, M., Panjanathan, R., Graceline Jasmine, S., Jani Anbarasi, L.: Diabetic retinopathy classification using CNN and hybrid deep convolutional neural networks. Symmetry 14(9) (2022). https://doi.org/10.3390/sym14091932
Marcus, G.: The next decade in AI: four steps towards robust artificial intelligence. arXiv:2002.06177 [cs.AI] (2020)
Ghnemat, R.: Hybrid framework for diabetic retinopathy stage measurement using convolutional neural network and fuzzy rules inference system. Appl. Syst. Innov. 5(102) (2022). https://doi.org/10.3390/asi5050102
https://www.kaggle.com/competitions/diabetic-retinopathy-detection/data. Accessed 12 Mar 2023
https://www.educba.com/keras-resnet50/. Accessed 12 Mar 2023
https://www.analyticsvidhya.com/blog/2021/06/transfer-learning-using-vgg16-in-pytorch/. Accessed 12 Mar 2023
https://www.geeksforgeeks.org/inception-v4-and-inception-resnets/. Accessed 12 Mat 2023
Simonyan, K., Vedaldi, A., Zisserman, A.: Deep inside convolutional networks: visualising image classification models and saliency maps. arXiv: 1312.6034v2 [cs.CV]
https://www.d-eyecare.com. Accessed 12 Mar 2023
Shen, Z., Yang, H., Zhang, S.: Neural network architecture beyond width and depth. arXiv: 2205.09459 [cs.LG], 14 January 2023
Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional neural networks. In: Advances in Neural Information Processing Systems 25, NIPS 2012 (2012)
Iandola, F., Han, S., Moskewicz, M., Ashraf, K., Dally, W., Keutzer, K.: SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5 MB model size. arXiv: 1602.07360 [cs.CV], 4 November 2016
Howard, A., et al.: MobileNets: efficient convolutional neural networks for mobile vision applications. arXiv: 1704.04861 [cs.CV], 17 April 2017
Moraga, C.: Introduction to fuzzy logic. Facta Universitas Series Electronics and Energetics, vol. 18, no. 2, pp. 319–328 (2005)
Katoch, S., Chauhan, S., Kumar, V.: A review on genetic algorithm: past, present and future. Multimedia Tools Appl. 80, 8091–8126 (2021)
Lin, C., Wu, K.: Development of revised ResNet-50 for diabetic retinopathy detection. BMC Bioinform. 24(157) (2023). https://doi.org/10.1186/s12859-023-05293-1
Acknowledgment
This work was supported by grant W/WI/4/2022 from Białystok University of Technology and funded with resources for research by the Ministry of Science and Higher Education in Poland.
The Author is grateful to experienced ophthalmologists who checked the results and provided their opinions about the processing pipeline and the marked regions within saliency maps.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Szymkowski, M. (2023). Neural Networks and Saliency Maps in Diabetic Retinopathy Diagnosis. In: Saeed, K., Dvorský, J., Nishiuchi, N., Fukumoto, M. (eds) Computer Information Systems and Industrial Management. CISIM 2023. Lecture Notes in Computer Science, vol 14164. Springer, Cham. https://doi.org/10.1007/978-3-031-42823-4_22
Download citation
DOI: https://doi.org/10.1007/978-3-031-42823-4_22
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-42822-7
Online ISBN: 978-3-031-42823-4
eBook Packages: Computer ScienceComputer Science (R0)