Radio Frequency Signal Based Human Silhouette Segmentation: A Sequential Diffusion Approach

Dataset

• Download the HIBER Dataset from HIBER ( Human Indoor Behavior Exclusive RF dataset ).

Conda Installation

# 1. Create a conda virtual environment.
conda create -n SDM python=3.6
conda activate SDM
conda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia

# 2. Clone the Repo and Install dependencies
git clone https://github.com/ph-w2000/SDM
pip install -r requirements.txt

Method

Training

This code supports multi-GPUs training.

python -m torch.distributed.launch --nproc_per_node=1 --master_port 48949 train.py -batch_size 8

• You can change the training hyperparameters in train.py file, such as dataset path, batch_size etc.
• To switch "WALK" dataset to "MULTI", you can change it in hiber_dataset.py file.

Inference

python -m torch.distributed.launch --nproc_per_node=1 --master_port 48949 test.py -batch_size 8

• You can change the training hyperparameters in test.py file, such as dataset path, batch_size etc.
• To switch "WALK" dataset to "MULTI", you can change it in hiber_dataset.py file.

Stage2

please switch to branch "fine-tune" get run stage2 code.

Citation

If you use the results and code for your research, please cite our paper:

@INPROCEEDINGS{10688347,
  author={Wen, Penghui and Hu, Kun and Yua, Dong and Ning, Zhiyuan and Li, Changyang and Wang, Zhiyong},
  booktitle={2024 IEEE International Conference on Multimedia and Expo (ICME)}, 
  title={Radio Frequency Signal based Human Silhouette Segmentation: A Sequential Diffusion Approach}, 
  year={2024},
  volume={},
  number={},
  pages={1-6},
  keywords={Wireless communication;Radio frequency;Wireless sensor networks;Motion segmentation;Dynamics;RF signals;Diffusion models;Wireless sensing;semantic segmentation;diffusion model;radio frequency},
  doi={10.1109/ICME57554.2024.10688347}}