Related Experiment Video
Updated: Apr 17, 2026
09:17
Using Retinal Imaging to Study Dementia
Published on: November 6, 2017
22.5K
A multiscale decomposition approach to detect abnormal vasculature in the optic disc
Carla Agurto1, Honggang Yu2, Victor Murray3
1Department of Electrical and Computer Engineering, University of New Mexico, Albuquerque, NM, USA; VisionQuest Biomedical LLC, Albuquerque, NM, USA.
Summary
This study introduces a new multiscale method for detecting neovascularization in the optic disc (NVD) from fundus images. The approach achieves high accuracy, aiding in early diagnosis of eye conditions.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Vision
Background:
- Neovascularization in the optic disc (NVD) is a critical indicator of various eye diseases.
- Accurate detection of NVD is essential for timely diagnosis and treatment.
Purpose of the Study:
- To develop and evaluate a multiscale method for automated NVD detection in fundus images.
- To improve the accuracy and efficiency of NVD identification.
Main Methods:
- Vessel segmentation using adaptive contrast enhancement and specialized techniques.
- Extraction of textural features via multiscale AM-FM, morphological granulometry, and fractal dimension.
- Classification using a linear Support Vector Machine (SVM) with 10-fold cross-validation.
Main Results:
- The method was evaluated on 300 fundus images.
- Achieved an Area Under the Curve (AUC) of 0.93.
- Reached a maximum classification accuracy of 88%.
Conclusions:
- The proposed multiscale method demonstrates high performance in detecting NVD.
- This technique offers a promising automated solution for NVD analysis in clinical settings.
