By Ya-Fang Hsieh, Kai-Lung Hua

This code is written in Python 3.8 and requires the packages listed in requirements.txt. Install with pip install -r requirements.txt preferably in a virtualenv.

mmWave-VitalSign/  
│  
├── CGU/  
│   └── src/  
│       ├── configs/  
│       ├── data_process/  
│       └── model/  
│  
└── datas/  
    ├── RobustVSDataset_anonymous/  
    └── cgu/  

• Baseline Dataset:
  The dataset for 'Rf vital sign sensing under free body movement' is self-collected. Detailed movement setup information is presented in the table below.

And related movement setup descriptions are presented in the table below.

• CGU Dataset:
  The CGU dataset includes 75 healthy participants. The related movement setup is presented in the table below.

In the CGU dataset, the training data includes two .csv files.

• move-all.csv contains motion data from 75 participants. Its fields include Tester, Motion, GT(HR), Radar(HR), Power, Distance. The field descriptions are as follows:

• fitness.csv contains the personalized fitness data of 75 participants. Its fields include Tester, Class, Score, and Gender. The field descriptions are as follows:

• The source code in folder CGU/src
• The dataset in foler datas/, include Baseline dataset and CGU dataset
• You don't need to change any dataset structure.

Direct to the CGU/src/configs folder

Both Baseline and CGU have their own config.yaml. In config.yaml, you can modify the relevant model parameters and output paths based on your needs.

• Baseline: baseline_config.yaml
• CGU: CGU_fitness_config.yaml

Navigate to the src folder and execute the command below. Once training is complete, the results will be saved in {The project path}/result/Baseline/ or {The project path}/result/CGU/. Based on the parameters set in {Dataset}_config.yaml, the saved results include the following items:

{Dataset}/excel: Evaluation results
{Dataset}/predict_result: Evaluation result images
{Dataset}/saved_model: Trained weights

• For Baseline train:
  python baseline_train.py

• For CGU train:
  python CGU_fitness_train.py

Evaluation metric : Mean Absolute Error (MAE) Heart rate error unit: beats per minute (BPM)

• Baseline:

• CGU:

Jian Gong, Xinyu Zhang, Kaixin Lin, Ju Ren, Yaoxue Zhang, and Wenxun Qiu, “Rf vital sign sensing under free body movement,” Proceedings of the ACM on Inter- active, Mobile, Wearable and Ubiquitous Technologies, vol. 5, no. 3, pp. 1–22, 2021.

Under review