Image Analysis for Markes-less Facial Tracking

Tracking of facial features is increasingly used in game and film industry as well as for other applications. Most tracking systems are currently using markers which unfortunately are tedious and cumbersome. Marker-less facial tracking is supposed to eliminate the disadvantages of the marker-based approaches. This thesis investigates different algorithms for marker-less tracking and presents how to apply them in a robust way. View-based and component sensitive normalized face images can achieve accurate tracking results based on the Active Appearance Algorithm. Post processing the parameters of global motions of the model smoothes the synthesized video sequence. Tacking results for faces and a tool developed for creating training images are also presented.


1. Introduction
1.1 Marker-based motion capture
1.2 Marker-less motion capture
1.3 Thesis motivation and purpose
1.4 Thesis overview
2. Facial Feature Tracking using Skin Texture
2.1 Material and calibration
2.2 Texture matching through 2D correlation
2.3 Alternative measurements of texture similarity
2.3.1 Mutual information criterion
2.3.2 filter & distance
2.4 Possible improvements
2.4.1 Gray level appearance
2.4.2 Derivative profile
2.5 Tracking experiments
2.6 Conclusions
3. Facial Motion Tracking using Active Face Model and Active Appearance Algorithm
3.1 Candide-3 face model
3.2 Face model adaptation
3.3 GSA model file
3.4 Normalized face model
3.4.1 Eigenface
3.4.2 Non-linear face space
3.4.3 Geometrical normalization
3.5 Image warping
3.5.1 Barycentric coordinate computation
3.5.2 Warping process
3.6 Texture synthesis
3.7 Facial tracking
3.8 Estimating the gradient matrix
3.9 Tracking experiments
3.10 Discussion
3.11 Possible improvements
4. View-based Texture Modeling
4.1 Capturing Setup
4.2 Modifications
4.2.1 Formula of model
4.2.2 Cropped frontal normalized face
4.2.3 Head pose estimation
4.2.4 Energy function
4.3 View-based texture modeling
4.4 Tracking experiments
4.5 Discussions
5. Component sensitive, post-processing, stereovision and real-time tracking
5.1 Component-based normalized face model
5.3 Tracking experiments
5.4 Capturing in stereo vision
5.5 Possibility of real-time tracking
5.6 Conclusions

Author: Wenlan, Yang

Source: Uppsala University Library

Download URL 2: Visit Now

Leave a Comment