In this repository we develop a framework for handling the multi-model data generated within the Emotion recognition for dementia project.
Goals and sub-goals are included in the issue-tracker of this repository.
Our overall goal is to build a system that can do the following:
- Load and store time series derived from multi-model data (video, audio, text) using software like OpenFace and Librosa.
- Make the derived data searchable, e.g. put it in a datbase and facilitate selection.
- Provide an interface for analytics and visualisation.