Face Biometric Applications Team Members: Faune Hughes, Daniel Lichter, Richard Oswald, and Michael Whitfield Clients: Fred Penna and Robert Zack
Project Overview A literature review of past and present face biometric systems was conducted. How anthropometrics relates to facial biometric systems The orbital (eye) region of 13 candidates was studied and analyzed. Two facial biometric systems were reviewed, Neurotechnology’s Verilook and Luxand’s FaceSDK, and experiments were conducted with the affects of aging in mind. A review of 360 Degree Web’s FACE software to determine how secure it is in securing portable device such as a laptop
Introduction
Identity theft has prompted the need for more secure applications Review of current technology to determine feasability of Facial Biometrics applications Experiments implementing facial biometrics to gain insight into problems of face aging.
Literature Review
Facial Biometrics What is it How is it used Benefits Literature Review
Dimensional Technology 2D vs. 3D Special Considerations Literature Review Facial Recognition by Grid Overcoming limitations
Live or Memorex Does the application care? or even notice? Human Validation Literature Review
Facial Biometrics in Use People Places Things Literature Review
Anthropometrics
Anthropometry = the study and measurement of human physical dimensions Pioneer in Anthropometry: Dr. Leslie Farkas Her defined “landmarks” prove that every face had different measurements Anthropometrics
Anthropometric Landmarks Anthropometrics
It is believed that the eye region does not change much over time. We measured the orbital region of each photo which consist of both the biocular distance and the intercanthal distance. Orbital Measurements
Anthropometrics Orbital Measurements
Methodology
Photo Database 44 photos from 19 subjects Digitized through webcam, digital camera, or scanner Anthropometrics Face Biometric Software Luxand’s FaceSDK Neurotechnology’s Verilook
Anthropometrics Mobile Face Biometric Software 360 Degree Web’s FACE software
Results Of Facial Recognition Experiments
Changes in False Acceptance Rates F.A.R. – margin of error
Photo Environment Attributes of the photo and purpose for which it was taken
Similarity Matrix
Gender and Ethnicity
Enrollment with Generalization Unique to Verilook Combines facial templates from multiple photos to give better matches
Summary Comparison between Luxand and Verilook showed both strengths and weakness Verilook merges all photos of single person into one; better matching Luxand shows better consistency of a person over a ten year span Future Work: It would be best to use a public face database because of it’s more controlled environment