(team 1)jackie abbazio, sasha perez, denise silva and robert tesoriero (team 2) faune hughes, daniel...
TRANSCRIPT
(Team 1)Jackie Abbazio, Sasha Perez, Denise Silva and Robert Tesoriero(Team 2) Faune Hughes, Daniel Lichter, Richard Oswald and Michael Whitfield
Clients: Fred Penna and Robert Zack
Literature Review Facial Anthropometrics
The orbital (eye) region of 13 candidates was studied and analyzed.
Software Selection: Neurotechnology’s Verilook, Luxand’s FaceSDK,
360 Degree Web’s FACE and others. Conduct Experiments with a focus on Security
and Aging.
Team 1 Primary Objectives: Conduct technology reviews of selected
facial recognition software. Select several facial recognition
applications for team 2 to perform experiments.
Review the state of Facial Biometric Technologies.
Team 2 Primary Objectives:
Experiments
Evaluate enrollment, data collection, feature extraction, classification, ease of use, performance, and other capabilities.
Primary Objectives of the Study:SoftwareSecurityAging ApplicationsStrengths and WeaknessesCosts, Benefits and Limitations
Objectives
2D vs. 3D Barriers and Obstacles Emerging Technologies False Acceptance Rates (FAR) False Rejection Rates (FRR)
Anthropometry is 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
Landmarks
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.
Enrolled into class databaseEnrolled into class database
This 1979 image matched with 2007 image2008 image with 51% similarity
This 1979 image matched with 2007 image2008 image with 51% similarity
Designed for biometric system developers and integrators.
Allows for easy integration and rapid development of biometric applications using functionality.
Can perform simultaneous multiple face detections with the ability to process 100,000 faces per second and it recommends the minimum image size to be 640x480 pixels .
The results show that the photo from 1969 matched a photo from 2008 with a similarity score of 18 or 10%. This result is comparable with the FaceSDK age
identification test, where the same image from 1969 matched the same photo from 2008 with a 61.9% similarity
rate.
VeriLook Identification and Authentication ResultsVeriLook Identification and Authentication Results FaceSDK Identification and
Authentication ResultsFaceSDK Identification and Authentication Results
Photo Database 44 photos from 19 subjects
Digitized through webcam, digital camera, or scanner
360 Degree Web’s Face
Attributes of the photo and purpose for which it was taken
Unique to Verilook Combines facial templates from multiple
photos to give better matches
All products tested have strengths and weaknesses. None suitable for security applications.
Verilook merges all photos of single person into one; better matching
Limited than Luxand. Only allows enrollment of high-res images.
Did not perform as well as FaceSDK software
Luxand works very well in identifying face similarity among
people in a group worked relatively well matching aged images
Future Work - Emerging Technologies