Computer scientists in Korea and Vietnam have teamed up to develop an algorithm that can tell if you're happy or sad using images from your computer's webcam.
Hyung-Il Choi of the School of Media at Soongsil University in Seoul, Korea, working with Nhan Thi Cao and An Hoa Ton-That of Vietnam National University in Ho Chi Minh City, built the algorithm with the objective of allowing video games to read and react to a player's emotions.
The system uses mathematical processing to analyse several indicators - eyebrow position, openness of the eyes, mouth shape and a handful of other factors. It can gather data from even low-resolution images - just 48 pixels square.
99% accuracy
It then correlates the data it receives with a list of human emotions - anger, disgust, fear, joy, sadness, surprise and a neutral expression. In tests on two sets of sample facial images, the researchers claim it was able to achieve 96.3% and 99.85% accuracy respectively.
The team says that the resulting data could be useful in video games - transferring the player's emotions to their avatar, for example, or activating appropriate actions to communicate with other players in different scenarios.
"When facial expressions recognition of players is applied in an intelligent game system, the experience can become more interactive, vivid and attractive," the team said, adding that the same system could be used to track emotional expressions of voice actors working on animated movies for more realistic animation.
The details of the system were published in the International Journal of Computational Vision and Robotics.
