MotionCue

QHack 2023 submission. MotionCue is a motion annotation solution that "transcribe" movement to text.

See more details in our Devpose: link

Architecture Design

Resources

Training our model

• https://opencv.org/
• https://scikit-learn.org/stable/modules/ensemble.html#forest
• https://google.github.io/mediapipe/solutions/pose

Inspecting our data

• Tensorflow pre-trained models - https://github.com/tensorflow/tfjs-models

Displaying to our ui

• https://github.com/u-wave/react-youtube
• https://mui.com/material-ui/getting-started/installation/
• tabs - https://reactcommunity.org/react-tabs/

Texts on pose detection

• https://www.americananthro.org/ImageDescriptions?navItemNumber=25126
• https://www.cs.ubc.ca/~lsigal/Publications/SigalEncyclopediaCVdraft.pdf