Related Experiment Video
Updated: Sep 20, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.0K
Investigations of CNN for Medical Image Analysis for Illness Prediction
K Nirmala1, K Saruladha1, Kenenisa Dekeba2
1Department of Computer Science and Engineering, Pondicherry Technological University, Puducherri, India.
Computational Intelligence and Neuroscience
|June 9, 2022
Summary
This study enhances diabetic retinopathy screening by using a CNN model to analyze retinal images. The approach improves diagnostic accuracy for exudates, aiding early detection and treatment.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Science
Background:
- Diabetic retinopathy exudates are common but challenging to diagnose from noisy images.
- Automated and computer-aided diagnosis systems are needed to overcome image quality issues.
Purpose of the Study:
- To improve the prediction rate for diabetic retinopathy analysis.
- To enhance accuracy, sensitivity, specificity, and prediction rates in diagnosing diabetic retinopathy.
Main Methods:
- Images were divided into patches of varying sizes and fed into a Convolutional Neural Network (CNN).
- A mathematical approach was used to analyze exudate characteristics (prevalence, shape, color, density) within image patches.
Main Results:
- The CNN model achieved 78% accuracy, 78.8% sensitivity, and 78.3% specificity.
- Positive and negative predictive values were also determined, indicating reliable diagnostic performance.
Conclusions:
- The proposed method offers a viable approach for automated diabetic retinopathy diagnosis using CNNs.
- Improved diagnostic metrics suggest potential for early detection and risk assessment, leading to timely hospital treatment.
