Track, cluster and classify people using the ti mmwave sensor.
This repo is a messy collection of the stuff i did in my thesis. The usefull parts are:
QTGui.py is a capture program for mmwave data.
use -h to see the options
labeler/labeler.py is a program to inspect the recorded data, and add labels to them.
RFClassifier.py is a script to try different classification methods based on scikit-learn.
recurrent_networks.ipynb is a jupyter notebook using tensorflow to classify the tracked objects. Implemented is a LSTM and GRU network. You can use this online in google colab: https://colab.research.google.com/drive/1w-dlOCp_RiF4aoebhJrN33Tra0lLX_z4?usp=sharing
Spread throughout are scripts for inspecting, plotting, and various tools to work with the mmwave data and intermediate formats.