Related Experiment Video
Updated: Jul 17, 2026
09:17
Using Retinal Imaging to Study Dementia
Published on: November 6, 2017
Retinal layer segmentation in OCT images with a 2.5D cross-slice feature fusion module for glaucoma assessment
Hyunwoo Kim1, Heesuk Kim2, Chaewon Lee3
1Department of Mechanical Engineering, Yonsei University, Seoul, 03722, Republic of Korea.
Biomedical Optics Express
|July 16, 2026
Summary
Accurate glaucoma diagnosis requires reliable retinal layer segmentation. Our novel 2.5D framework with cross-slice feature fusion improves segmentation accuracy and robustness, balancing context and efficiency for clinical use.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Vision
Background:
- Accurate glaucoma diagnosis and monitoring depend on precise retinal layer segmentation in Optical Coherence Tomography (OCT) images.
- Existing 2D methods lack inter-slice context, leading to inconsistencies, while 3D methods are computationally demanding.
Purpose of the Study:
- To develop a computationally efficient 2.5D framework for robust retinal layer segmentation in OCT images.
- To improve the accuracy and consistency of segmentation across adjacent B-scans.
Main Methods:
- A novel 2.5D segmentation framework integrating a U-Net-like architecture.
- Implementation of a cross-slice feature fusion (CFF) module to capture inter-slice contextual information.
Main Results:
- The proposed framework demonstrated improved segmentation accuracy and robustness on clinical and public datasets.
- Achieved an 8.56% reduction in mean absolute distance and a 13.92% reduction in root mean square error compared to methods without CFF.
Conclusions:
- The 2.5D framework effectively balances contextual awareness with computational efficiency for retinal layer segmentation.
- Enables anatomically reliable delineation for automated glaucoma evaluation and potential clinical applications.
