Related Experiment Video
Updated: Jul 13, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.8K
H-Deep-Net: A deep hybrid network with stationary wavelet packet transforms for Retinal detachment classification
Sonal Yadav1, R Murugan1, Tripti Goel1
1Bio-Medical Imaging Laboratory (BIOMIL), Department of Electronics and Communication Engineering, National Institute Of Technology Silchar, Assam-788010, India.
Medical Engineering & Physics
|October 14, 2023
Summary
A new hybrid AI model accurately detects retinal detachment (RD) using fundus images. This automated approach offers high sensitivity and specificity, aiding early diagnosis and treatment of this severe eye condition.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Automated disease diagnosis is crucial for managing a growing global population.
- Retinal detachment (RD) is a severe, acute ocular illness requiring prompt diagnosis.
- Early detection of RD is essential to minimize vision loss and improve patient outcomes.
Purpose of the Study:
- To develop an automated, hybrid AI model for the early detection of retinal detachment (RD).
- To enhance diagnostic accuracy and speed for RD using retinal fundus images.
Main Methods:
- A hybrid approach combining best basis stationary wavelet packet transform for image analysis.
- Utilizing modified VGG19-Bidirectional long short-term memory for deep feature extraction.
- Employing Adaptive Boosting technique for classification of extracted features.
Main Results:
- The proposed model achieved high performance metrics: 99.67% sensitivity, 95.95% specificity, 98.21% accuracy, 97.43% precision, 98.54% F1-score, and 0.9985 AUC.
- The model demonstrated effective detection of retinal detachment on a current dataset.
- Results indicate the potential for further enhancement with larger datasets.
Conclusions:
- The developed hybrid AI model provides an accurate and efficient automated method for retinal detachment diagnosis.
- This approach can assist ophthalmologists in timely identification and treatment of RD patients.
- The model shows significant promise for clinical application in diagnosing ocular diseases.
