Spring 2002
CIS 6930-002 Image Based Biometrics
Time: 12:30pm to 1:45pm, Mondays and Wednesdays
Office Hours: 2pm to 3pm (Monday and Wednesday) (or any other time mutually agreed upon)
Location: ENG 003 (First of classes), ENB 313 (Rest of the semester)

Facilitator: Sudeep Sarkar

 
 

Goal: To investigate pattern recognition and computer vision techniques that can be used to identify humans from image data, be it video sequences or still frames.  What are the limitations of these techniques? How do you evaluate their effectiveness? Can information from different imaging modalities be combined to arrive at better performance? Can you identify humans at a distance? What are the ethical implications of this technology?

Impact on society: This technology has obvious impact on automated surveillance tasks. It can also be used to authenticate person to allow access of information remotely, e.g. web based financial transactions, automated teller machines, access to secure area or equipment, and so on.

Pre-reqs: Digitial Image Processing (CAP 5400) or Computer Vision (CAP 6415)

Scoring of Student Performance will be based on progress on project, presentations made to the class, and class participation. There will be no formal exams. There might be a few quizes. Tentatively:

       45% project (30% report + 15% presentation),

       15% discussion,

       20% Core technique notes,

       10% class participation,

       10% worksheets, project status reporting, project proposal etc.

Textbook: There is no standard textbook. Relevant papers will be provided or made available. A couple of reference texts: (i) Biometrics (A Jain, R Bolle, and A Pankanti Eds) (ii) Pattern Classification (Duda, Hart, and Stork)

Some Relevant Links

The National Biometrics Consortium, which serves as the US Government's focal point for research, development, test, evaluation, and application of biometric-based personal identification/verification technology.

The DARPA Human ID at a Distance program

Data Sets Available

Deadlines


Tentative Course Schedule
 

WEEK 

TOPICS

Notes

Jan 7, Jan 9

Intro to Biometrics, 
Description of available data sets, baseline performances, and possible projects

 

Jan 14, Jan 16

Gait baseline algorithm and performances, CMC, ROC

 

Jan 21, Jan 23

MLK Day, 
Core Technique 1: Bayesian decision theory

Project Proposal due on 1/23 

Jan 28, Jan 30

Core Technique 1: Bayesian decision theory

Topic paper list due on 1/28

Feb 4, Feb 6

Core Technique 2: Discriminant surfaces, Multivariate Gaussian densities

 

Feb  11, Feb 13

Core Technique 3: Maximum likelihood estimation

 

Feb  18, Feb 20

Core Technique 3: MLE with mixture models,

Core Technique 4: Expected Maximization

 

Feb 25, Feb 27

Core Technique 4: Expected Maximization

 

Mar 4, Mar 6

Core Technique 5: Hidden Markov Models

 

Mar 11, Mar 13

 Spring Break

 

Mar 18

Core Technique 6: Bayesian networks

Core technique – first draft due

March 20

Project Updates (5 minutes each, 3 slides atmost: what, how, any results, plans)

 

Mar 25

Topic Discussions: Fingerprint, Hand (Zongyi)

 

Mar 27

Topic Discussions: Iris (Ayush), Retina (Sorin)

 

Apr 1

Topic Discussions: Infrared (Lakshmi), Gait (Isidro)

 

Apr 3

Topic Discussions: Ear (Yan), Face

 

Apr 8,

Topic Discussions: Multimodal biometrics (Paddu)

Core technique – revised version due

Apr 10

Project Presentations: Isidro, Zongyi

 

Apr 15

Project Presentations: Ayush, Sorin

 

Apr 17 

Project Presentations: Yan, Yong

 

Apr 22,

Project Presentations: Paddu, Lakshmi

 

Apr 24

Ethical Implication of Biometric Technology

Project Final Report Due