Related Experiment Video
Updated: Dec 7, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.2K
Automated Quality Assessment and Image Selection of Ultra-Widefield Fluorescein Angiography Images through Deep
Henry H Li1,2, Joseph R Abraham1, Duriye Damla Sevgi1
1The Tony and Leona Campane Center for Excellence in Image-Guided Surgery and Advanced Imaging Research, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA.
Translational Vision Science & Technology
|September 30, 2020
Summary
This study shows a deep learning system can automatically classify ultra-widefield fluorescein angiography (UWFA) image quality. This technology may reduce manual grading and improve image selection for clinical use.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Ultra-widefield fluorescein angiography (UWFA) generates numerous images with variable quality.
- Manual assessment of UWFA image quality is time-consuming and subjective.
Purpose of the Study:
- To evaluate the feasibility of a deep learning-based automated system for classifying UWFA image quality.
- To assess the system's accuracy in distinguishing between gradable and ungradable images.
Main Methods:
- A deep learning model was trained on 3543 UWFA images graded by experts into four quality categories.
- Model performance was evaluated on two independent test sets, including a randomly selected set and a balanced set of high/low quality images.
Main Results:
- The automated system achieved high accuracy in classifying gradable versus ungradable images (89.0% and 94.0% on test sets).
- High sensitivity (90.5%-98.6%) and specificity (87.0%-81.5%) were observed.
- Area under the curve (AUC) for two-class classification was strong (0.920 and 0.980).
Conclusions:
- Deep learning enables feasible and accurate automated classification of UWFA image quality.
- This system can potentially reduce manual grading workload and provide real-time feedback.
- Clinical implementation could streamline image review and improve diagnostic efficiency.
