Towards Large-scale Single-shot Millimeter-wave Imaging for Low-cost Security Inspection

1. System requirements

1.1 All software dependencies and operating systems

The project has been tested on Ubuntu 22.04.1 LTS.

1.2 Versions the software has been tested on

The project has been tested on Python version >= 3.8.

1.3 Any required non standard hardware

There is no non-standard hardware required for this project.

2. Installation guide

2.1 Download the repository

To install the software, clone the repository and run the following command in the terminal:

git clone https://github.com/bianlab/MMW.git

2.2 Install the requirements

To install the requirements, make sure that you have installed Anaconda, and run the following command in the terminal:

conda create -n mmw python=3.8
conda activate mmw

pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
pip install tqdm scipy

3. Demo

3.1 Untrained learning reconstruction

We have provided some test cases in the Reconstruction/ExemplarData folder. You can run the following command in the terminal:

cd ./Reconstruction
python Main.py --data_folder ExemplarData/ --data_name data_0.25 --n_iter 1000 --use_init True

The reconstructed 3D data (.mat) is in the Reconstruction/Result folder.

3.2 Concealed target detection

We have provided some test cases in the Detection/images/0.1 and Detection/images/0.25 folders corresponding to 10% sampling ratio reconstructions and 25% sampling ratio reconstructions. There are two detection networks in the Detection/models folder, one for 10% sampling ratio, and the other for 25% sampling ratio. Make sure you have yolov8 package or run the command:

pip install ultralytics

You can run the following command in the terminal:

cd ./Detection
yolo predict imgsz=640 model=models/0.1.pt source=images/0.1 name=0.1 show_conf=False iou=0.5
yolo predict imgsz=640 model=models/0.25.pt source=images/0.25 name=0.25 show_conf=False iou=0.5

The detection images are in the Detection/runs/detect/0.1 or Detection/runs/detect/0.25 folder.

4. Instructions for use

4.1 How to run the reconstruction code on your data

We have provided the sparse pattern (.mat) in the Mask folder. You can make your own input data for untrained learning by multiplying the mask by the echoes. The input data (.mat) consists of 4 parts:

S_echo_real_sampling: real part of the sparsely sampled echo

S_echo_imag_sampling: imaginary part of the sparsely sampled echo

sampling_mask: sparse pattern

scene_init: initial reconstruction by RMA or other method (optional)

You can run the following command in the terminal:

python Main.py --data_folder your_folder --data_name you_own_data --n_iter iteration_number --use_init True

4.1 How to run the detection code on your data

Put your images in the Detection/images/0.1 or Detection/images/0.25 folders. Then run the following command in the terminal:

yolo predict imgsz=640 model=models/0.1.pt source=images/0.1 name=0.1 show_conf=False iou=0.5
yolo predict imgsz=640 model=models/0.25.pt source=images/0.25 name=0.25 show_conf=False iou=0.5

The code will check both the folders and produce the detection results.