ryankemmer / falldetection Goto Github PK

This repository is a collection of deep learning models created to detect potentially life threatening falls in videos. Models are written in python and utilize tensorflow, pandas and numpy. The most successful model uses the popular 'OpenPose' library to perform feature extraction of humans in videos, and then uses a CNN/LSTM framework to predict is a person is experiencing a potentially fatal fall.