Related Experiment Video
Updated: Sep 5, 2025
12:49
A Method for 3D Reconstruction and Virtual Reality Analysis of Glial and Neuronal Cells
Published on: September 28, 2019
12.9K
A Pixel-Based Machine-Learning Model For Three-Dimensional Reconstruction of Vitreous Anatomy
Alan Thi1, K Bailey Freund1,2, Michael Engelbert1,2
1Vitreous Retina Macula Consultants of New York, New York, NY, USA.
Translational Vision Science & Technology
|July 8, 2022
Summary
A new machine learning model reconstructs 3D vitreous anatomy from swept-source optical coherence tomography scans. This enables detailed visualization of vitreous structures and degeneration, aiding in understanding eye diseases.
Area of Science:
- Ophthalmology
- Medical Imaging
- Machine Learning
Background:
- The vitreous humor's complex 3D structure is challenging to visualize non-invasively.
- Advanced imaging techniques are needed to study vitreous anatomy and pathology.
Purpose of the Study:
- To develop a machine learning model for 3D reconstruction of vitreous anatomy using swept-source optical coherence tomography (SS-OCT).
Main Methods:
- SS-OCT imaging of healthy subjects.
- Image processing using Fiji (ImageJ) with Trainable Weka Segmentation.
- 3D volume rendering using TomViz.
Main Results:
- Developed a machine learning model for 3D vitreous reconstruction from SS-OCT data.
- Successfully visualized clinically relevant vitreous features in 3D.
- Demonstrated high-resolution 3D visualization of vitreous anatomy and degeneration.
Conclusions:
- The developed model enables comprehensive 3D examination of vitreous structure.
- Potential applications for diseases like vitreomacular traction and proliferative diabetic retinopathy.
