Advanced Detection of Diabetic Retinopathy: Employing Hybrid Deep Neural Networks and Multi-Scale Image Analysis Techniques DOI Open Access

A. M. Mutawa,

G. R. Hemalakshmi,

N. B. Prakash

et al.

Published: Jan. 30, 2024

Diabetic retinopathy (DR), a major complication of prolonged diabetes, poses significant risk vision loss. Early detection is critical for effective treatment, yet traditional diagnostic methods by ophthalmologists are time-consuming, costly, and subject to variability. This study introduces novel approach employing hybrid Convolutional Neural Network-Radial Basis Function (CNN-RBF) classifier integrated with Multi-Scale Discriminative Robust Local Binary Pattern (MS-DRLBP) features enhanced DR detection. We implemented advanced image preprocessing techniques, including noise reduction, morphological operations, Otsu’s thresholding, optimize blood vessel segmentation from retinal images. Our method demonstrates exceptional performance in screening DR, achieving an average 96.10% precision, 95.35% sensitivity, 97.06% specificity, accuracy. These results significantly outperform offer promising tool remote efficient DR. Applied publicly available datasets, this research contributes the development accessible, accurate ophthalmology, potentially reducing global burden diabetic

Language: Английский

Advanced Detection of Diabetic Retinopathy: Employing Hybrid Deep Neural Networks and Multi-Scale Image Analysis Techniques DOI Open Access

A. M. Mutawa,

G. R. Hemalakshmi,

N. B. Prakash

et al.

Published: Jan. 30, 2024

Diabetic retinopathy (DR), a major complication of prolonged diabetes, poses significant risk vision loss. Early detection is critical for effective treatment, yet traditional diagnostic methods by ophthalmologists are time-consuming, costly, and subject to variability. This study introduces novel approach employing hybrid Convolutional Neural Network-Radial Basis Function (CNN-RBF) classifier integrated with Multi-Scale Discriminative Robust Local Binary Pattern (MS-DRLBP) features enhanced DR detection. We implemented advanced image preprocessing techniques, including noise reduction, morphological operations, Otsu’s thresholding, optimize blood vessel segmentation from retinal images. Our method demonstrates exceptional performance in screening DR, achieving an average 96.10% precision, 95.35% sensitivity, 97.06% specificity, accuracy. These results significantly outperform offer promising tool remote efficient DR. Applied publicly available datasets, this research contributes the development accessible, accurate ophthalmology, potentially reducing global burden diabetic

Language: Английский

Citations

1