Facial Expression Analysis (Emotion Analysis)

The technology on Rose Research’s platform uses state of the art machine learning techniques, capable of recognizing the basic facial expressions of emotion (based on the work of Dr. Paul Ekman, these expressions are Neutral, Happy, Sad, Angry, Surprised, Scared, Disgusted)

The technology on our platform does this using 3 steps:

Face Detection
The technology on our platform uses the popular Viola-Jones algorithm to detect the location and size of faces (including eye detection).

Face Modeling
The technology also builds an accurate model of a face based on the Active Appearance model described by Coots & Taylor (using more than 500 key points on the face).

Face Classification
Achieved by feeding a deep neural network more than 10,000 annotated images.

Studies employing this technology can be conducted online using a computer webcam and paired with eye-tracking technology so that we can understand exactly what the respondent is observing while we detect changes in the user’s facial expressions.