Citation

Raza, Asif and Musa, Shahrulniza and Khalid, Ahmad Shahrafidz and Alam, Muhammad Mansoor and Su’ud, Mazliham Mohd and Noor, Fouzia (2025) Multiclass Diabetic Retinopathy: Hybrid Metaheuristic Particle Swarm Optimization and Classification for Severity Grading and Feature Extraction. Engineering, Technology & Applied Science Research, 15 (6). pp. 30317-30323. ISSN 2241-4487

Text Microsoft Word - ETASR_V15_N6_pp-30317-30323.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

This study aimed to effectively classify a Diabetic Retinopathy (DR) image dataset consisting of colored images by developing a hybrid and robust transfer learning model called MOB-PSO, integrating MobileNet with Particle Swarm Optimization (PSO) to enhance performance and accuracy. A wellstructured dataset is crucial for building a high-performance model capable of accurate feature extraction and precise identification of image features within each class. Traditional statistical algorithms often struggle to classify colored images accurately, leading to errors in detecting diseases within DR image datasets. To reduce error rates and improve classification accuracy, this study developed a hybrid, reliable, and optimized image classification model. The DR dataset consists of ten distinct classes. Experimental results demonstrate that the MOB-PSO model surpasses state-of-the-art algorithms in terms of accuracy, robustness, precision, recall, and F1 score, achieving optimal validation loss values. Specifically, the MOBPSO model recorded training and validation losses of 0.1515 and 0.1853, respectively, with corresponding accuracies of 98.58% and 96.7%. The precision, recall, and F1-score were 0.9744, 0.9657, and 0.9698, respectively, showcasing the model's effectiveness.

Item Type:	Article
Uncontrolled Keywords:	MobileNet, CNN, deep learning
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Computing and Informatics (FCI)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	06 Feb 2026 09:01
Last Modified:	06 Feb 2026 09:01
URII:	http://shdl.mmu.edu.my/id/eprint/15215

Downloads

Edit (login required)