Related Experiment Video
Updated: Dec 22, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.2K
Multi-view deep learning for rigid gas permeable lens base curve fitting based on Pentacam images
Sara Hashemi1, Hadi Veisi2, Ebrahim Jafarzadehpur3
1Faculty of New Sciences and Technologies, University of Tehran, Tehran, Iran.
Medical & Biological Engineering & Computing
|May 5, 2020
Summary
This study introduces a deep learning method using Pentacam images to quickly identify rigid gas permeable (RGP) lens base curves for irregular astigmatism patients. This approach enhances fitting accuracy and reduces patient chair time.
Area of Science:
- Ophthalmology
- Computer Science
- Medical Imaging
Background:
- Fitting rigid gas permeable (RGP) lenses for irregular astigmatism presents a significant clinical challenge.
- Current fitting methods, while accurate, lack high-pace solutions for expert assistance.
Purpose of the Study:
- To develop a deep learning solution for rapid identification of RGP lens base curves using Pentacam refractive maps.
- To improve the efficiency and accuracy of RGP lens fitting for patients with irregular astigmatism.
Main Methods:
- Utilized a dataset of 247 Pentacam four refractive maps for multi-view corneal structure analysis.
- Employed scratch-based and transfer learning Convolutional Neural Network (CNN) architectures (AlexNet, GoogLeNet, ResNet) for feature extraction.
- Implemented a feature fusion technique to aggregate extracted information.
Main Results:
- The multi-view scratch-based CNN achieved an R-squared value of 0.849, comparable to existing methods.
- Transfer learning approaches demonstrated superior performance over scratch-based CNN models.
- CNNs applied to multi-view Pentacam images enabled fast RGP lens base curve detection.
Conclusions:
- Deep learning, specifically CNNs on multi-view Pentacam images, offers a fast and accurate solution for RGP lens base curve identification.
- This technology has the potential to increase patient satisfaction and reduce clinical chair time in RGP lens fitting.
- The proposed methodology provides a valuable tool for eye care professionals managing irregular astigmatism.
