IMO

This is official Pytorch implementation of "Joint Segmentation and Grading with Iterative Optimization for Multimodal Glaucoma Diagnosis"

@inproceedings{Wang2026JointSA,
  title={Joint Segmentation and Grading with Iterative Optimization for Multimodal Glaucoma Diagnosis},
  author={Zhiwei Wang and Yuxing Li and Meilu Zhu and Defeng He and Edmund Y. Lam},
  year={2026},
  url={https://api.semanticscholar.org/CorpusID:286572762}
}

Framework

Recommended Environment

• torch 1.13.1
• cudatoolkit 11.8
• torchvision 0.14.0
• mmcv 2.2.1
• mmcv-full 1.7.2
• mmsegmentation 0.30.0
• numpy 1.26.4
• opencv-python 4.10.0.84

Experiments

Dataset & Checkpoints & Results

The checkpoints and results can be in IMO. Download MSRS dataset from GAMMA. If you need to evaluate other datasets, please organize them as follows:

├── /dataset
    GAMMA/
    ├── Cls
    │   ├── train
    │   │   ├── class1
    │   │   ├── class2
    │   │   └── class3
    │   └── val
    │       ├── class1
    │       ├── class2
    │       └── class3
    └── Seg
        ├── class_label.txt
        |
        ├── test
        │   ├── label
        │   ├── oct
        │   └── vi
        │   
        └── train
            ├── label
            ├── oct
            └── vi
        ......

Evaluate model

python

python test_model.py

run sample

python

python test_demo.py --img="./images/00131D_vi.png" --ir="./images/00131D_ir.png" --checkpoint="./exps/Done/msrs_vi_ir_meanstd_ConvNext_fusioncomplex_8083/best.pth" --segout="./seg.png"

To Train

Before training IMO, you need to download the GAMMA dataset and putting it in ./datasets.

Then running python

python train_model.py

Segmentation comparison

Grading comparison

If this work is helpful to you, please cite it as：

@inproceedings{Wang2026JointSA,
  title={Joint Segmentation and Grading with Iterative Optimization for Multimodal Glaucoma Diagnosis},
  author={Zhiwei Wang and Yuxing Li and Meilu Zhu and Defeng He and Edmund Y. Lam},
  year={2026},
  url={https://api.semanticscholar.org/CorpusID:286572762}
}

Acknowledgements