/SelfAffectiveMemory

Describes the self-organizing affective memory model: https://ieeexplore.ieee.org/abstract/document/7965832

Primary LanguagePythonGNU Lesser General Public License v2.1LGPL-2.1

Affective Memory

This repository holds the code for the Affective Memory: A self-organizing and online model that learns specific characteristics of individual's facial expressions.

Train Model

The Affective Memory is composed of two modules: a deep neural network, based on the FaceChannel neural network.

The Affective Memory is implemented as a Growing-When-Required network that learns online how to map the arousals and valences from the Face Channel into a temporally-correlated reading.

Webcam Demo

Screenshot

The demo displays the current arousal/valence output coming from the FaceChannel, and the neurons learned by the affective memory. Each neuron is represented by a dot, and the brighter dot represents the newest neurons. The red dot represents the average arousal and valence readings.

Check the video.

Requirements

Install all the libraries on the requirements.txt file.

Related Publications

License

All the examples in this repository are distributed under the Creative Commons CC BY-NC-SA 3.0 DE license. If you use this corpus, you have to agree with the following itens:

  • To cite our reference in any of your publication that make any use of these examples. The references are provided at the end of this page.
  • To use this model for research purpose only.

Contact

Pablo Barros - pablo.alvesdebarros@iit.it