We propose a System for Processing In-situ Bio-signal Data for Emotion Recognition and Sensing in the form factor of glasses, that combines customized modular bio-signal acquisition circuit with a suite of algorithm and machine-learning based classifiers to produce accurate bio-signals and emotion classes. The proposed platform uses data from noncontact sensors to provide: 1) eye & eyebrow detection, 2) pupillometry, 3) zygomaticus muscle (smile muscle) movements, and 4) head movements. The platform can also be extended to include functionalities provided by contact-based sensors, such as 5) heart rate and 6) 8-channel EEG.
This folder contains all the PCB designs for the main board and power potential board.
The pipeline folder contains the code for data preprocessing, facial landmark detection, facial expression detection, and emotion classification. #GUI This folder includes the materials for the preliminary experiment MATLAB GUI.