/human-activity-recognizer

Primary LanguagePython

human-activity-recognizer

this is skeleton based human activity recognizer Using Tiny-YOLO oneclass to detect each person in the frame and use AlphaPose(resnet50) to get skeleton-pose and then use ST-GCN model to predict action from every 30 frames of each person tracks.

suported actions: Standing, Walking, Sitting, Lying Down, Stand up, Sit down, Fall Down.

Prerequisites

  • Python > 3.7
  • Pytorch > 1.7
  • opencv-python > 4.5.1
  • tqdm > 4.56
  • scipy > 1.6
  • matplotlib > 3.3.4

Pre-Trained Models

models download this file and extrect to project directory

Basic Use

  • use web cam as a input
python det.py -C
  • pyqt based app
python app.py
  • input video as a input
python det.py -V cs4.mp4
  • run in colab without install any dependencies Untitled10.ipynb

Reference