Purpose
This package offers classes and methods for pose estimation on images. This can be done either with deep learning based facial landmark detection or by detecting ArUco markers.
Installation
Get the latest published version: pip install headpose
or install directly from GitHub: pip install git+https://github.com/OleBialas/headpose.git
To use landmark detection you additionally have to install
pytorch and torchvision
Example
import cv2
from headpose.detect import PoseEstimator
est = PoseEstimator() #load the model
# take an image using the webcam (alternatively, you could load an image)
cam = cv2.VideoCapture(0)
for i in range(cv2.CAP_PROP_FRAME_COUNT):
cam.grab()
ret, image = cam.retrieve()
image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
cam.release()
est.detect_landmarks(image, plot=True) # plot the result of landmark detection
roll, pitch, yawn = est.pose_from_image(image) # estimate the head pose
Sources & Further Reading
This blog post nicely explained the concepts and mathematics behind pose estimation and this tutorial walks through the single steps of detecting facial landmarks with pytorch