Related Experiment Video
Updated: Jul 8, 2026
12:28
Quantification of Vascular Parameters in Whole Mount Retinas of Mice with Non-Proliferative and Proliferative Retinopathies
Published on: March 12, 2022
3.5K
Quantum chimp-enanced SqueezeNet for precise diabetic retinopathy classification.
Anas Bilal1,2, Muhammad Shafiq3, Waeal J Obidallah4
1College of Information Science and Technology, Hainan Normal University, Haikou, 571158, China.
Scientific Reports
|April 15, 2025
Summary
Early detection of diabetic retinopathy (DR) is crucial to prevent blindness. This study introduces a hybrid Quantum Chimp Optimization Algorithm (QCOA) and SqueezeNet model for highly accurate DR classification, improving patient outcomes.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Diabetic retinopathy (DR) is a leading cause of preventable blindness globally.
- Prolonged hyperglycemia damages retinal blood vessels, necessitating early detection for intervention.
- Current diagnostic methods require enhancement for improved accuracy and efficiency.
Purpose of the Study:
- To develop and validate an advanced hybrid approach for enhanced diabetic retinopathy classification.
- To improve the accuracy, sensitivity, and specificity of DR detection using artificial intelligence.
- To facilitate earlier diagnosis and intervention for diabetic retinopathy to prevent vision loss.
Main Methods:
- A hybrid model integrating Quantum Chimp Optimization Algorithm (QCOA) with SqueezeNet for feature extraction and classification.
- SqueezeNet efficiently extracts critical features from segmented fundus images with low computational cost.
- QCOA optimizes Support Vector Machine (SVM) parameters and performs feature selection for refined classification.
Main Results:
- The hybrid QCOA-SqueezeNet-SVM model achieved exceptional classification accuracy of 99.80%.
- The system demonstrated high sensitivity (99.90%) and perfect specificity (100%) in DR detection.
- The approach significantly enhanced SVM performance, optimizing the classification model.
Conclusions:
- The proposed hybrid approach offers a highly accurate and efficient method for diabetic retinopathy classification.
- The integration of QCOA and SqueezeNet shows significant potential for improving early DR detection rates.
- Clinical implementation of this AI-driven system can lead to better patient outcomes by enabling timely treatment.
