Skip to content

Latest commit

 

History

History
33 lines (29 loc) · 1.01 KB

README.md

File metadata and controls

33 lines (29 loc) · 1.01 KB

Real-Time Head Pose Estimation

  • A PyTorch implementation of real-time head pose estimation.
  • The original work can be found here.
  • However, instead of using dlib, mtcnn is implemented to detect faces.
  • Download the pretrained model here from my Google drive.
  • This is part of my summer research project at Monash University.

Quickstart: Run Inference in Real Time

python3 run_inference.py --pretrained [PRETRAINED_MODEL]

Offline Testing

python3 test_on_video.py --pretrained [PRETRAINED_MODEL] --video [VIDEO] --output_string [NAME] --fps 30

Requirements

  • numpy
  • pytorch
  • torchvision
  • opencv2
  • pillow
  • mtcnn
pip3 install mtcnn

Credits