Face Recognition and Tracking from Video.
| Sponsor |
Irfan Essa
irfan@cc.gatech.edu
320 CoC
|
| Area |
Intelligent Systems / GVU |
Problem
Face Recognition from video is getting extremely popular these days
and many systems claiming very high accuracies have appeared in the
recent years. The purpose of this project to study an existing system,
read and understand some of the seminal papers in this area and propose
methods to enhance and improve these methods.
Here is what you need to do.
-
Read the follwoing papers on face recognition.
-
Turk, M., and Pentland, A., "Face Recognition Using
Eigenfaces", In Proceedings of IEEE Conference on Computer Vision
and Pattern Recognition, 1991.
-
Study an implementation of this system and other systems that we have
on our machines in the college (One of my students will show you how).
-
Add more images to the database and experiment.
-
Read the papers that deal with head and face orientation
-
Trevor Darrell, Irfan Essa, and Alex Pentland "Task-specific
Gesture Analysis in Real-Time using Interpolated Views", In IEEE
Transactions on PAMI January 1997.
-
Michael Black and Yaser Yacoob, "Tracking and Recognizing Rigid
and Non-Rigid Facial Motions using Local Parametric Model of Image
Motion". In Proceedings of International Conference on Computer
Vision, 1995.
-
Based on the above two methods, propose a method for adding the
functionality in the above mentioned systems to allow for head/face
orientation.
Background
Deliverables
-
Write a 3-5 page Report addressing the following issues:
-
Give a brief overview of what you understand is the face recognition
problem. Is is tractable?
-
Explain the method proposed in Turk and Pentland Paper.
-
Describe the implementation in the systems you tried out.
-
Describe its limitations
-
Propose an improvement. Defend your proposal as to how you would do it
and why would it work.
Evaluation
Based on the report turned in to the sponsor of the project by the due
date.
updated by tucker, 9/7/97, 5:45pm.