PowerSkel: A Device-Free Human Skeleton Estimation Framework for Electric Power Operation Based on CSI Signal

Training

• The example data is in the ./dataset/csi directory. The data involves power station environment and is collected in cooperation with the power department, the data is confidential. If you need cooperation, please email us. If you want to replace your own data, you can make changes according to the form from the given csv file in ./dataset/csi. Email address: cunyiyin1125@gmail.com

• python train.py --root ./dataset/csi/ --model_names conti conti --num_workers 8 --print_freq 10 --gpu-id 0,1,2,3

Requirements

Our environment and package versions are:

• python 3.7.16
• pytorch 1.13.1
• 4 NVIDIA RTX A6000 GPUs with 48G memory per GPU

Acknowledgements

The code has been published in the IEEE Internet of Things Journal and is available for reading and citation： Yin, Cunyi, et al. "PowerSkel: A Device-Free Framework Using CSI Signal for Human Skeleton Estimation in Power Station." IEEE Internet of Things Journal (2024).

This repo is partly based on the following repos, thank the authors a lot.

• shaoeric/Online-Knowledge-Distillation-via-Collaborative-Learning
• gpeyre/SinkhornAutoDiff