CBP Biometric Fusion and Multimodal Systems 3

Question 1

Bayesian fusion in multimodal biometrics computes the final decision by:

Accepted Answer

Combining likelihood ratios using Bayes' theorem to estimate posterior probability of identity

Answer

Averaging the match scores from all enrolled templates

Answer

Selecting the matcher with the highest individual EER

Answer

Normalizing all feature vectors to unit length before matching

Question 2

How does multimodal biometric fusion help address the Failure to Enroll (FTE) problem?

Accepted Answer

It allows enrollment to succeed even if one modality fails, using remaining modalities

Answer

It eliminates FTE entirely by using redundant sensors

Answer

It forces users to re-enroll in all modalities simultaneously

Answer

FTE is not affected by multimodal systems

Question 3

In a weighted sum fusion rule, weights are typically assigned based on:

Accepted Answer

The Equal Error Rate (EER) or other performance measures of each individual matcher

Answer

The physical size of the biometric sensor

Answer

The order in which matchers were deployed in the system

Answer

The geographic location of the biometric capture

Question 4

Anti-spoofing (liveness detection) in a multimodal system is best described as:

Accepted Answer

An additional layer that verifies sample authenticity before or after fusion

Answer

A replacement for fusion algorithms

Answer

Applicable only to fingerprint modalities

Answer

A form of feature-level fusion

Question 5

What is the concept of 'adaptive fusion' in multimodal biometric systems?

Accepted Answer

Dynamically adjusting fusion weights or strategy based on real-time signal quality or context

Answer

Using only one modality and adapting its threshold based on usage history

Answer

Replacing one biometric modality with another when hardware fails

Answer

Applying the same fixed rule across all users and conditions

Question 6

Which statement best describes rank-level fusion in multimodal biometrics?

Accepted Answer

It consolidates ranked candidate lists from multiple matchers into a single ranked list

Answer

It combines binary decisions from each matcher using Boolean logic

Answer

It merges raw biometric feature vectors into a single combined template

Answer

It averages normalized match scores across all matchers

Question 7

The NIST Multiple Biometric Grand Challenge (MBGC) primarily evaluated:

Accepted Answer

Fusion of face and iris biometrics from video and still imagery under challenging conditions

Answer

Encryption standards for biometric template storage

Answer

Fingerprint interoperability across different sensor types

Answer

Behavioral biometrics for continuous authentication