(2003) Biometrics Consortium Conference - Securing Restricted Site
Upcoming SlideShare
Loading in...5
×
 

(2003) Biometrics Consortium Conference - Securing Restricted Site

on

  • 376 views

This study evaluated the performance of a commercially available face recognition algorithm for the verification of an individual's identity across three illumination levels. The lack of research ...

This study evaluated the performance of a commercially available face recognition algorithm for the verification of an individual's identity across three illumination levels. The lack of research related to lighting conditions and face recognition was the driver for this evaluation. This evaluation examined the influence of variations in illumination levels on the performance of a face recognition algorithm, specifically with respect to factors of: age, gender, ethnicity, facial characteristics, and facial obstructions.

Statistics

Views

Total Views
376
Views on SlideShare
376
Embed Views
0

Actions

Likes
0
Downloads
6
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

(2003) Biometrics Consortium Conference - Securing Restricted Site (2003) Biometrics Consortium Conference - Securing Restricted Site Presentation Transcript

  • Securing a Restricted Site Biometric Authentication at Entry Point Eric Kukula & Stephen Elliott, Ph.D. epkukula@tech.purdue.edu & sjelliott@tech.purdue.edu Biometric Standards, Performance, and Assurance Laboratory www.biotown.purdue.edu © Purdue University 2006 Department of Industrial Technology, School of Technology, Purdue University, West Lafayette, IN 47906 Module 1 1
  • Introduction • This study evaluated the performance of a commercially available face recognition algorithm for the verification of an individual’s identity across three illumination levels • The lack of research related to lighting conditions and face recognition was the driver for this evaluation • This evaluation examined the influence of variations in illumination levels on the performance of a face recognition algorithm, specifically with respect to factors of: – Age, gender, ethnicity, facial characteristics, and facial obstructions © Purdue University 2006 Module 1 2
  • Experimental Setup • This evaluation took place in Biometric Standards, Performance and Assurance Laboratory in the School of Technology at Purdue University. Evaluation Area © Purdue University 2006 Module 1 3
  • Light • This study evaluated the performance of a commercially available face recognition algorithm capabilities of face recognition in three illumination levels. • The first light level, 7 -12 illuminance (lux) referred to as low light, was determined by logging 60 minutes of data from a local campus restaurant. • The second light level, 800 – 815 illuminance (lux) referred to as high light, was determined by logging 60 minutes of data from the Industrial Technology office. © Purdue University 2006 • The third light level, 407 – 415 illuminance (lux), referred to as medium light, was determined by taking the mean of the other two light levels. Module 1 4
  • Sample Images from the Data Set Low Light Medium Light High Light 8 Lux 423 Lux 800 Lux © Purdue University 2006 Module 1 5
  • Volunteer Crew Gender Male 73% Female 27% Age 20 - 29 73% 30 - 39 10% 40 - 49 10% 50 & older 7% Ethnicity African American 0% Asian/Pacific Islander 13% Caucasian 73% Hispanic 13% Native American 0% Features Facial Hair 30% Glasses 24% © Purdue University 2006 Hair over Face 0% Hats 6% Scars 3% Module 1 None 36% 6
  • Testing Protocol • Each enrollment collected a 100 images of the participants face from different directions to form a template – Subjects enrolled three times 1. Low light (7 – 12 lux) 2. Medium light (407 -412 lux) 3. High light (800 – 815 lux) • Each participant made three verification attempts in each light scenarios over three visits for a total of 27 verification attempts Evaluation Matrix Enrollment Conditions 7 - 12 lux 407 - 412 lux 800 - 815 lux © Purdue University 2006 3 Low 3 Low 3 Low Verification Attempts 3 Medium 3 Medium 3 Medium 3 High 3 High 3 High Module 1 7
  • Results Failure to Enroll Rates & Failure to Acquire Rates FTE* FTA Low Light (7 - 12 lux) 6.25% 0.92% Medium Light (407 - 412 lux) 9.09% 0.65% High Light (800 - 815 lux) 3.22% 0.00% * FTE included subjects that wore hats. The testing protocol placed no restrictions on participants’ style of dress. After 3 attempts failed, the subject was asked to remove the hat. When removed, the subject was able to enroll. Verification Rates Low Medium High Enrollment Enrollment Enrollment 7 - 12 lux 407 - 412 lux 800 - 815 lux Verification Attempts © Purdue University 2006 7-12 lux 89.62% 73.88% 80.55% Verification Attempts 407 - 412 lux 57.40% 91.48% 89.44% Verification Attempts Module 1 800 - 815 lux 58.70% 95.37% 94.25% 8
  • Conclusions • The results of this study show that there are still significant challenges with regard to illumination levels and face recognition especially at lower light levels • Further research will be conducted at Purdue University’s Biometric Standards, Performance and Assurance Laboratory dealing with lighting effects and background changes • This report may be downloaded at http://www.tech.purdue.edu/it/resources/bi © Purdue University 2006 ometrics/bc2003.htm Module 1 9