Related Experiment Video
Updated: Nov 25, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
DeepRetina: Layer Segmentation of Retina in OCT Images Using Deep Learning
Qiaoliang Li1, Shiyu Li1, Zhuoying He1
1National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, Department of Biomedical Engineering, School of Medicine, Shenzhen University, Xueyuan Avenue, Nanshan District, Shenzhen, Guangdong Province, China.
DeepRetina, a novel deep neural network, automates retinal layer segmentation from optical coherence tomography (OCT) images. This method achieves high accuracy, aiding in early disease diagnosis and improving clinical efficiency.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Accurate segmentation of retinal layers in optical coherence tomography (OCT) images is crucial for diagnosing and monitoring various fundus retinal diseases.
- Manual segmentation is time-consuming and subjective, necessitating automated solutions.
Purpose of the Study:
- To develop and validate DeepRetina, a deep neural network-based method for automating the segmentation of retinal layers in OCT images.
- To assess the performance of DeepRetina in terms of accuracy and efficiency for clinical applications.
Main Methods:
- DeepRetina utilizes an improved Xception65 model for feature extraction, combined with an atrous spatial pyramid pooling module for multiscale feature analysis.
- An encoder-decoder module refines feature maps to accurately delineate retinal layer boundaries.
- The method was validated on a dataset of 280 retinal OCT volumes (40 B-scans per volume).
Main Results:
- DeepRetina demonstrated excellent performance with a mean intersection over union (IoU) of 90.41% and sensitivity (Se) of 92.15%.
- Segmentation accuracy for all individual retinal layers, including the nerve fiber layer and pigment epithelium layer, exceeded 88% IoU and Se.
- The automated segmentation significantly reduces the need for manual segmentation, enhancing work efficiency.
Conclusions:
- DeepRetina provides an effective automated solution for retinal layer segmentation in OCT images.
- The method shows significant potential for early diagnosis of fundus retinal diseases and can be adapted for segmenting other tissue types.
- Automated segmentation improves diagnostic capabilities and clinical workflow efficiency.
