MM-Net: 5th Place Solution to HUAWEI PRCV Challenge 2021 Alzheimer's Disease Classification Task

This repo contains the supported pytorch code and configuration files to reproduce alzheimer's disease classification results of MM-Net. Official website of the competition. Link to our team's Huawei homepage.

Environment

Prepare an environment with python=3.6, and then run the command "pip install -r requirements.txt" for the dependencies.

Data Preparation

• For experiments we used one dataset:

  • Clinical sMRI: https://competition.huaweicloud.com/information/1000041489/circumstance?zhishi=

• File structure

    train_data
    |--- train
    |   |--- Subject_xxxx.npy
    |   |--- Subject_xxxx.npy
    |   |--- ...
    |   |--- train_open.csv
    MM-Net
    |---model.py
    |---customize_service.py
    |---std.npy
    |---mean.npy
    |---pip-requirements.txt
    ...
  

Pre-Trained Base Model For PRCV Challenge

• AD-CLS: https://marketplace.huaweicloud.com/markets/aihub/modelhub/detail/?id=18ab4679-279c-4f41-af64-3e90ec583fdf
• Download AD-CLS pre-trained model and add it under MM-Net folder before running test.py

Train/Test

The entries of this competition are deployed on Huawei Cloud to run and test, and if you want to run locally, you need to modify the inference code.

• Train : Run the train script on PRCV 2021 Training Dataset with Base model Configurations.

python model.py --train_url your_path --data_url your_data_path

• Test : Run the test script on PRCV 2021 Training Dataset.

python customize_service.py 

Acknowledgements

Thanks to HUAWEI Cloud for providing the competition platform.