An evolutionary supply chain management service model based on deep learning features for automated glaucoma detection using fundus images

Glaucoma, a multifaceted eye condition, poses a high risk of vision impairment. Initially, most automated approaches segment the primary system and assess the clinical measurements to classify and screen for glaucoma. The proposed customized convolutional neural network (CNN) model for automated glaucoma detection, built using deep learning techniques, can assist many stakeholders in the supply chain management network. These stakeholders may include eye hospitals, healthcare service providers, doctors, ophthalmologists, patients, insurance companies, etc. The deployed model comprises four learnable layers, i.e., three convolution layers and a flattened layer. The customized CNN model learned the deep features with the least number of tunable parameters. Subsequently, a combined feature reduction strategy called principal component analysis (PCA) and linear discriminant analysis (LDA) to reduce the dimensions of feature sets. Finally, a classification is carried out by utilizing an extreme learning machine (ELM). The hidden node parameters of ELM are optimized with the help of the modified particle swarm optimization (MOD-PSO) technique. The generalized performance of the proposed model has been enhanced by employing 5-fold stratified cross-validation. The proposed model deployed on two standard datasets, G1020 and ORIGA. The experimental results show that the proposed computer-aided diagnosis (CAD) model achieves an accuracy of 97.80% and 98.46% on the G1020 and ORIGA datasets, respectively. The customized CNN model outperforms as compared to other state-of-the-art models with a significantly less number of features and could help the decision-makers of supply chain management networks.