A novel Adaptive Neural Network-Based Laplacian of Gaussian (AnLoG) classification algorithm for detecting diabetic retinopathy with colour retinal fundus images

• Published in: Neural computing & applications 2024-03, Vol.36 (7), p.3513-3524
• Main Authors: Ramasamy, Manjula Devi, Periasamy, Keerthika, Periasamy, Suresh, Muthusamy, Suresh, Ramamoorthi, Ponarun, Thangavel, Gunasekaran, Sekaran, Sreejith, Sadasivuni, Kishor Kumar, Geetha, Mithra
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial Intelligence; Classification; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Data Mining and Knowledge Discovery; Diabetes; Diabetic retinopathy; Edge detection; Eye diseases; Image Processing and Computer Vision; Machine learning; Medical imaging; Neural networks; Original Article; Probability and Statistics in Computer Science
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-023-09324-z

Diabetic retinopathy (DR) is a human eye disease in which the eye’s retina is damaged in diabetics. Diabetic retinopathy can be diagnosed by manually interpreting retinal fundus images, even though that takes longer to diagnose. Among these, the most challenging task in diagnosing the DR disease is edge detection in retinal fundus images to identify the region of infection and its severity. This paper aims to use the adaptive neural network-based Laplacian of Gaussian (AnLoG) classification algorithm on features extracted from diverse retinal fundus images to improve DR disease diagnostic accuracy and reduce training time. Based on the retinal fundus image in the Messidor dataset, the consequence of the proposed AnLoG classification algorithm for detecting diabetic retinopathy is compared to traditional supervised BPN machine learning algorithms and other contemporary techniques. AnLoG has proved its supremacy in terms of accuracy (97.29%), recall (94.64%), precision (93.13%), and F-Score (93.80%). Simulation results show that the proposed technique performs well compared to the existing approach.