Another technique we use is Buckshot (Cutting et al.pdf

D Face Modeling from Perspective
Views and Contour
Based Generic
Model Kamsum LEE
Kinhong WONG
Siuhang OR
Yiufai FUNG Department of Computer Science and Engineering The Chinese University of Hong Kong email fkslee
khwong
shor
yunggcsecuhkeduhk Abstract Threedimensional human head modeling is useful in videoconferencing or other vir tual reality applications
However manual construction of D models using CAD tools is often expensive and timeconsuming
Here we present a robust and ecient method for the construction of a D human head model from perspective images viewing from dierent angles
In our system A generic head model is rst used then three images of the head are required to adjust the deformable contours on the generic model to make it more closer to the target head
Our contributions are as follows
Our system uses per spective images that are more realistic than orthographic projection approximation used in earlier works
Also for shaping and positioning face organs we present a method for estimating the camera focal length and the D coordinates of facial landmarks when the camera transformation is known
We also provide an alternative for the D coor dinates estimation using epipolar geometry when the extrinsic parameters are absent
Our experiments demonstrated that our approach produces good and realistic results
Keywords facial images face model model reconstruction facial landmarks per spective projection generic model epipolar geometry essential matrix Introduction Realistic modeling of D objects is essential to many multimedia applications
For example in verylowbitrate video compression modelbased image coding is one of the most eective techniques
In this method the users face is rst modeled ge ometrically
Then the D model is transmitted to the receiver at the beginning of a  communication session
During the communication only the movement of the user is transmitted such as rotation and