Scribd
Upload
52 views

National Institute of Technology Rourkela

This document contains a summary of a biometric security exam containing multiple choice and written answer questions. The exam covers topics like biometric authentication, fingerprint classification systems, fingerprint matching algorithms, iris recognition, and performance evaluation metrics. It contains 4 sections with a total of 15 sub-questions assessing concepts such as cancelable biometrics, fingerprint matching time calculations, Henry classification, local orientation in fingerprints, and evaluation metrics like rank-based identification and equal error rate.

Uploaded by

amulya
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
52 views

National Institute of Technology Rourkela

This document contains a summary of a biometric security exam containing multiple choice and written answer questions. The exam covers topics like biometric authentication, fingerprint classification systems, fingerprint matching algorithms, iris recognition, and performance evaluation metrics. It contains 4 sections with a total of 15 sub-questions assessing concepts such as cancelable biometrics, fingerprint matching time calculations, Henry classification, local orientation in fingerprints, and evaluation metrics like rank-based identification and equal error rate.

Uploaded by

amulya
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA

 Biometric Security CS6217  End Sem Autumn 2020  B.Tech. Sem 8  Pg 1  FM 50  2 Hrs
 Answer ALL questions.  All parts of a question MUST be answered together.  Mere answers without proper
explanation will not fetch marks.  Variation in quality of answer will vary the secured marks.

1. a. Differentiate biometric authentication, surveillance, and forensic identification?


b. Find the center of a square with length 5 px using distance transform. Clearly show interim steps.
c. What is cancelable biometrics? [1.5+2.5+1]

2. Consider the following data experimentally obtained from a fingerprint biometric system which stores the image
of enrolled subjects as template: Acquisition and localization time = 1.5 ms, Feature extraction time = 2.5 ms,
Time to fetch a located template from database = 2 ms, 1:1 matching time = 3 ms, all operations are sequential.
Number of subjects enrolled in system = 150. Any other operation consumes negligible time.
a. What is matching time for verification when an enrolled subject gives his fingerprint?
b. What is the time for enrolling a new subject’s fingerprint in database?
c. What is the matching time for identification when an enrolled subject gives his fingerprint?
d. How much the matching time for identification increases if the number of enrolled subjects become 300 (other
parameters remaining same)? Is this increment in time static / linear / quadratic / exponential?
e. What happens when a non-enrolled subject gives his fingerprint in verification / identification mode? What is
the response time in this case?
f. If the system stores features (rather than storing templates), what is the matching time for identification when
an enrolled subject gives his fingerprint?
g. If an enrolled subject takes 200 ms to provide his claimed identity (e.g. typing his ID number) and 5ms to
place his finger (all tasks are sequential), for what criteria of population, the biometric system would be faster in
identification mode than verification mode? [1.5+1.5+1.5+(2+2)+1.5+2.5+2.5]

3. a. In a 10-finger Henry classification system, find the primary classification number of a subject having: loop in
right middle and left little finger, whorl in right thumb and ring fingers and left index and middle finger, arches
in right index finger.
b. Explain how to find local orientation θ(i, j) at a pixel (i, j) in a intensity normalized fingerprint image.
c. Define crossing number. Draw sample 3 × 3 pixel grids for explaining crossing numbers 0 to 4 and their
corresponding properties.
d. What is a latent fingerprint, inked fingerprint, and scanned fingerprint? [5+3+4+3]

a. Consider the 5 × 5 matrix depicting similarity based match score of 5  


subjects stored in database with 5 queries (for each of the subjects). Find 12 8 3 2 2
the d0 index for this score distribution. 0 7 2 4 1
 
7 5 14 4 2 
b. What is integrodifferential operator and how it is used? Explain  
8 1 4 2 3
transformation of coordinates for iris normalization.
c. What is meant by ‘a Rank-7 bound identification system with 55 enrolled 6 3 9 2 15
4. subjects’ ? What is the value of cumulative Rank-7 identification probability
in this system?
d. Can accuracy curve be reconstructed from ROC curve? Can EER be
found from ROC curve?
e. What is Hamming distance? Is it a similarity based score or dissimilarity
based score?
[3.5+(2.5+2.5)+3.5+(1+1)+1]

 BEST OF LUCK 

You might also like