ljl199711129/WiGr

WiFi-based Cross-Domain Gesture Recognition via Modified Prototypical Networks

WiFi-based Cross-Domain Gesture Recognition via Modified Prototypical Networks

About the code files

models (folder)

These are the different types of Dual-PAth PN

• LSTM_CSI.py: the network is based on CNN and LSTM (Model_type_1).

• MobiV3_CSI_model.py: the network is based on the 2D convolutional network (Model_type_2).

• ResNet_CSI_model.py: the network is based on the 1D convolutional network (Model_type_3) (Using in WiGr).

• Prototypical_CnnLstmNet.py: the pytorch-lightning version of Model_type_1;

• Prototypical_2DMobileNet.py: the pytorch-lightning version of Model_type_2;

• Prototypical_1DResNet.py: the pytorch-lightning version of Model_type_3 (Using in WiGr).

reimplement (folder)

Reimplementation of the related models: Widar3.0, EI, JADA, SignFi, ARIL, WiAG

Training a Dual-Path Prototypical Network

Install dependencies

• This code has been tested on Ubuntu 16.04 with Python 3.6 and PyTorch-1.8.0.
• Install PyTorch and torchvision.
• Install (pytorch-lightning)[https://github.com/PyTorchLightning/pytorch-lightning]

Download the datasets

• Download Widar3.0 dataset: http://tns.thss.tsinghua.edu.cn/widar3.0/
• Download ARIL dataset: https://github.com/geekfeiw/ARIL
• Download CSIDA dataset: https://pan.baidu.com/s/1Teb8hVWDxhOw0aIoVnS7Qw Password:lwp6

Train and Test the model

• Run python in_domain_run.py. This will run in-domain training and place the results into lightning_logs (this folder will be automatic constructed).
• Run python cross_domain_run.py. This will run cross-domain training
• the parameter_config.py is the configurations of the cross-domain and the in-domain experiments