Emotion Recognition OpenCV
I made this as a summer research project. This emotion recognition program recognizes people's emotions solely through facial movement by using an algorithm known as dense optical flow.
It then takes the dense optical flow data, compact it in a matrix, and feeds it into a machine learning algorithm, support vector machines, which trains it so that it would learn the facial movement patterns, look at new videos of people's emotions, and correctly determine the person's emotion.
EDIT: The level of confidence display is a bit buggy and will be working on that later. But overall, it can predict the correct emotion at an 80-90 percent accuracy range.
I made this as a summer research project. This emotion recognition program recognizes people’s emotions solely through facial movement by using an algorithm known as dense optical flow.
It then takes the dense optical flow data, compact it in a matrix, and feeds it into a machine learning algorithm, support vector machines, which trains it so that it would learn the facial movement patterns, look at new videos of people’s emotions, and correctly determine the person’s emotion.
EDIT: The level of confidence display is a bit buggy and will be working on that later. But overall, it can predict the correct emotion at an 80-90 percent accuracy range.
