The University of Sheffield
Department of Computer Science

Wenyi Hu Undergraduate Dissertation 2014/15

Kinect-based facial motion capture

Supervised by S.Maddock

Abstract

Nowadays, enterainment has been seamlessly intergrated with everyone's daily life. There are plenty of things we do to relax ourselves in leisure time, playing video games and watching movies must be the highest scored ones. However, conventional video games and movies can no longer satisfy most of us as we require more virtual impacts and interations. With the help of computer facial animation and Kinect sensor, the game is able to track the player's face as well as changing facial expressions through the Kinect sensor. Then the characters' facial motions will change with the players in the game. The games thus become more interactive with players and create an attractive virtual reality for the player.

The amin of this project is to capture the Kinect based facial motions, and reflect the changing facial expressions on a 2D sample cartoon face.

In this paper, 2D morphing algorithm is applied to morph a source image using the changed marked lines only. The presentation of the result is generally in effect. But our program would occasionally present strange expressions when marked lines intertwine. In future, we will carry on to avoid the interwoven situation of marked lines, hence to improve the results of cartoon facial changes.