Presentation is loading. Please wait.

Presentation is loading. Please wait.

Improving ATM Security via Facial Recognition CPSC510 James Maxlow November 25 th, 2002.

Similar presentations


Presentation on theme: "Improving ATM Security via Facial Recognition CPSC510 James Maxlow November 25 th, 2002."— Presentation transcript:

1 Improving ATM Security via Facial Recognition CPSC510 James Maxlow November 25 th, 2002

2 Proposal  Cameras in use at automatic teller machines should take still images of users  A facial recognition scheme should be added to the software used to verify users at ATMs  This scheme should match a picture of the user at the ATM with a picture of the account holder in the bank’s database 2 of 11

3 Reasoning  ATM fraud costs U.S. banks an average of $15,000 each year… hundreds of millions in total  This cost is borne by bank customers  Current ATM validation schemes are limited to access cards and PINs  Card theft, PIN theft and cracking, stealing of account information by bank employees all contribute to fraud schemes 3 of 11

4 Algorithm  Take customer’s picture(s) when account is opened and allow user to set non- verified transaction limits  At ATM, use access card and PIN to pre- verify user  Take user’s picture, attempt to match it to database image(s)  If match is successful, allow transaction  If match is unsuccessful, limit available transactions 4 of 11

5 Can ATMs Support This?  Most current generation ATMs run Windows CE, 2000, XP Embedded, or Linux – these machines can run facial recognition software locally ATM Bank Computer 1: Image, account no, PIN 2: Account no, PIN 3: Bank-held customer image 4: Processing 5: User verified 5 of 11

6 Can ATMs Support This?  Older ATMs run DOS or OS/2 – these machines can offload the processing to the bank’s computers ATM Bank Computer 1: Image, account no, PIN 2: Image, account no, PIN 4: User verified message 5: User verified message 3: Processing 6 of 11

7 Is Facial Recognition Reliable?  Matching a user image with any image in a database… facial evaluation… is still problematic – digital airport screenings, public surveillance cameras, etc. have high error rates  But this isn’t a problem here!  We must only match a user image with one known image from a database… facial verification… this technique has low error rates under good conditions (3% - 10%) 7 of 11

8 What Variables Affect Verification? 1. Lighting 2. Angle of view 3. Extreme facial expressions 4. Facial hair 5. Glasses Please follow these guidelines to help us verify your identity: Face the camera, holding still until you hear the beep Maintain a normal facial expression If you are wearing glasses, please remove them 8 of 11

9 What Is The Best Verification?  Local Feature Analysis  This method analyzes the geometric relationships between facial features  In some versions it builds a 3-D topography of the user’s face and uses this topography for later comparisons 9 of 11

10 What Would I Actually Do?  Find open-source LFA recognition programs  Write an ATM black box module  Create two databases of images  Tweak and test recognition programs with ATM module and images  Rewrite ATM module into client/server version with encryption to emulate ATM/bank interactions  Add USB camera control to client 10 of 11

11 Summary  Access card / PIN provides insufficient ATM security  Adding facial verification to the process can greatly decrease fraudulent transactions  Current ATMs have the power to perform verification locally given a software change  I will create a simulated system using a custom ATM black box and an open-source LFA recognition program 11 of 11


Download ppt "Improving ATM Security via Facial Recognition CPSC510 James Maxlow November 25 th, 2002."

Similar presentations


Ads by Google