Vol.:(0123456789) 1 3 Journal of Ambient Intelligence and Humanized Computing https://doi.org/10.1007/s12652-020-01963-7 ORIGINAL RESEARCH Deep neural networks to predict diabetic retinopathy Thippa Reddy Gadekallu 1  · Neelu Khare 1  · Sweta Bhattacharya 1  · Saurabh Singh 2  · Praveen Kumar Reddy Maddikunta 1  · Gautam Srivastava 3,4,5 Received: 18 January 2020 / Accepted: 6 April 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020 Abstract Diabetic retinopathy is a prominent cause of blindness among elderly people and has become a global medical problem over the last few decades. There are several scientifc and medical approaches to screen and detect this disease, but most of the detection is done using retinal fungal imaging. The present study uses principal component analysis based deep neural network model using Grey Wolf Optimization (GWO) algorithm to classify the extracted features of diabetic retinopathy dataset. The use of GWO enables to choose optimal parameters for training the DNN model. The steps involved in this paper include standardization of the diabetic retinopathy dataset using a standardscaler normalization method, followed by dimen- sionality reduction using PCA, then choosing of optimal hyper parameters by GWO and fnally training of the dataset using a DNN model. The proposed model is evaluated based on the performance measures namely accuracy, recall, sensitivity and specifcity. The model is further compared with the traditional machine learning algorithms—support vector machine (SVM), Naive Bayes Classifer, Decision Tree and XGBoost. The results show that the proposed model ofers better performance compared to the aforementioned algorithms. Keywords Diabetic retinopathy · Deep neural networks · Dimensionality reduction · Principal component analysis · Grey Wolf Optimizer (GWO) · Deep learning 1 Introduction Diabetic retinopathy (DR) is a disease that damages the retina due to complications in diabetic mellitus leading to permanent damage of the eyes and sometimes even vision loss. This type of complication among patients has high pri- oritized chances of patients becoming blind if left untreated. As per the statistical data of the International Diabetes Fed- eration, there existed almost 425 million adults living with diabetes in 2017 and the number is going to increase pro- gressively to an estimated 629 million by 2045. There is also another group of 325 million with the risk of Type II diabetes in 2017 and the numbers of these patients are also progressively increasing throughout the world. Most people in this risk category belong to the age group between 40 and 59 years, wherein 1 out of 2 among 212 mil- lion people are completely ignorant and uninformed of their disease. Hence it is quite evident that diabetic retinopathy has the probability to soon become a signifcant health issue throughout the world. Obesity, unhealthy diet, and physi- cal inactivity are the primary factors responsible for Type 2 diabetes. However, it is essential to understand that diabetic * Gautam Srivastava srivastavag@brandonu.ca Thippa Reddy Gadekallu thippareddy.g@vit.ac.in Neelu Khare neelu.khare@vit.ac.in Sweta Bhattacharya sweta.b@vit.ac.in Saurabh Singh saurabh89@dongguk.edu Praveen Kumar Reddy Maddikunta praveenkumarreddy@vit.ac.in 1 Vellore Institute of Technology, Vellore, Tamil Nadu, India 2 Department of Industrial & Systems Engineering, College of Engineering, Dongguk University, Seoul 04620, Republic of Korea 3 Department of Mathematics and Computer Science, Brandon University, Brandon R7A 6A9, Canada 4 Research Center for Interneural Computing, China Medical University, Taichung 40402, Taiwan, Republic of China 5 College of Information and Electrical Engineering, Asia University, Taichung, Taiwan