Multi-Scale GCN-Assisted Two-Stage Network for Joint Segmentation of Retinal Layers and Disc in Peripapillary OCT Images


Data: Upon request

The codes are implemented in PyTorch and trained on NVIDIA Tesla V100 GPUs.


An accurate and automated tissue segmentation algorithm for retinal optical coherence tomography (OCT) images is crucial for the diagnosis of glaucoma. However, due to the presence of the optic disc, the anatomical structure of the peripapillary region of the retina is complicated and is challenging for segmentation. To address this issue, we developed a novel graph convolutional network (GCN)-assisted two-stage framework to simultaneously label the nine retinal layers and the optic disc. Specifically, a multi-scale global reasoning module is inserted between the encoder and decoder of a U-shape neural network to exploit anatomical prior knowledge and perform spatial reasoning. We conducted experiments on human peripapillary retinal OCT images. The Dice score of the proposed segmentation network is 0.820 ± 0.001 and the pixel accuracy is 0.830 ± 0.002, both of which outperform those from other state-of-the-art techniques.


  1. Collected dataset: Download our collected dataset at this link (it will be available soon!!!).
  2. Public dataset: Duke SD-OCT dataset
Train and test

Run the following script to train and test the two-stage model.

python --name tsmgunet -d ./data/dataset --batch-size 1 --epoch 50 --lr 0.001


Results on the collected dataset

Results on the public dataset

For more results, please refer to our [paper](


If you use the codes or collected dataset for your research, please cite the following papers:

author = {Jiaxuan Li and Peiyao Jin and Jianfeng Zhu and Haidong Zou and Xun Xu and Min Tang and Minwen Zhou and Yu Gan and Jiangnan He and Yuye Ling and Yikai Su},
journal = {Biomed. Opt. Express},
number = {4},
pages = {2204--2220},
title = {Multi-scale GCN-assisted two-stage network for joint segmentation of retinal layers and discs in peripapillary OCT images},
volume = {12},
year = {2021},
url = {},
doi = {10.1364/BOE.417212},


The codes are built on AI-Challenger-Retinal-Edema-Segmentation and GloRe. We sincerely appreciate the authors for sharing their codes.


If you have any questions, please contact