This project is still in progress. It aims to detect facial expressions from camera and apply it to an artificial model which is graphically controlled by muscles naturally inspired from human facial. The program is capable of applying any texture to artificial facial model and one of the target features is to add an option that the program could grab the texture from a live camera and apply it simultaneously to the artificial model. However, this option is not a priority for now.
Different methods are considered for detecting facial behaviour in computer vision from bags of features to deep learning but the application is still in development and the final version is not released. Currently I am detecting the face and then use ASM and STASM to get the facial feature point. I will keep all the computer vision methods I used for detecting expressions so that they could be compared with each other later. Not only the expressions are important for me in this project but the facial rotation and orientation is important as well and they should be applied to the artificial model.
The whole application is developed with C++, the only libraries used are MFC for GUI interface, OpenGL for graphical parts and OpenCV for vision algorithms and video IO. No extra graphical engine is used. The program is implemented under Visual studio and the use of MFC is restricted it to the windows. However, you should not face any problem if you launches the program under WINE environment in Linux. The binaries for the program are not available yet but they will be after the computer vision portion is completed.
I would like to thanks to my colleague who cooperate me in developing graphics portions of the application.
Following you can find descriptions of some features and tools available in the program through the screenshots of the program. click on one the images and navigate the screenshot gallery with descriptions available for each screenshot.