Computer vision · Notebook
Body Language Detection
A computer-vision experiment comparing landmark-based classification pipelines.
Overview
A computer-vision experiment comparing landmark-based classification pipelines.
Built notebooks for collecting labeled landmark data, training classifiers and exporting models.
The problem
Explore how video landmarks can be collected and used as classifier inputs.
My contribution
Built notebooks for collecting labeled landmark data, training classifiers and exporting models.
PythonMediaPipeOpenCVscikit-learnTensorFlow
Inside the build
- Collect labeled landmark data from webcam or recorded video.
- Compare scikit-learn pipelines and a TensorFlow Lite export.
Project notes
Archived research experiment
Body Language Detection is part of my work in computer vision · notebook. The links above provide the available project and implementation details.