ISO/IEC 19795-2:2007/Amd 1:2015

INTERNATIONAL ISO/IEC
STANDARD 19795-2
First edition
2007-02-01
AMENDMENT 1
2015-04-01
Information technology — Biometric
performance testing and reporting —
Part 2:
Testing methodologies for technology
and scenario evaluation
AMENDMENT 1: Testing of multimodal
biometric implementations
Technologies de l’information — Essais et rapports de performance
biométriques —
Partie 2: Méthodologies d’essai pour l’évaluation des technologies
et du scénario
AMENDEMENT 1: Essais des mises en oeuvre biométriques multimodales
Reference number
ISO/IEC 19795-2:2007/Amd.1:2015(E)
©
ISO/IEC 2015
ISO/IEC 19795-2:2007/Amd.1:2015(E)

© ISO/IEC 2015, Published in Switzerland
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form
or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior
written permission. Permission can be requested from either ISO at the address below or ISO’s member body in the country of
the requester.
ISO copyright office
Ch. de Blandonnet 8 • CP 401
CH-1214 Vernier, Geneva, Switzerland
Tel. +41 22 749 01 11
Fax +41 22 749 09 47
copyright@iso.org
www.iso.org
ii © ISO/IEC 2015 – All rights reserved

ISO/IEC 19795-2:2007/Amd.1:2015(E)

Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards
bodies (ISO member bodies). The work of preparing International Standards is normally carried out
through ISO technical committees. Each member body interested in a subject for which a technical
committee has been established has the right to be represented on that committee. International
organizations, governmental and non-governmental, in liaison with ISO, also take part in the work.
ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of
electrotechnical standardization.
The procedures used to develop this document and those intended for its further maintenance are
described in the ISO/IEC Directives, Part 1. In particular the different approval criteria needed for the
different types of ISO documents should be noted. This document was drafted in accordance with the
editorial rules of the ISO/IEC Directives, Part 2 (see www.iso.org/directives).
Attention is drawn to the possibility that some of the elements of this document may be the subject of
patent rights. ISO shall not be held responsible for identifying any or all such patent rights. Details of any
patent rights identified during the development of the document will be in the Introduction and/or on
the ISO list of patent declarations received (see www.iso.org/patents).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation on the meaning of ISO specific terms and expressions related to conformity
assessment, as well as information about ISO’s adherence to the WTO principles in the Technical Barriers
to Trade (TBT), see the following URL: Foreword — Supplementary information.
The committee responsible for this document is ISO/IEC JTC 1, Information technology, Subcommittee
SC 37, Biometrics.
© ISO/IEC 2015 – All rights reserved iii

ISO/IEC 19795-2:2007/Amd.1:2015(E)

Introduction
Testing and reporting methods in this part of ISO/IEC 19795 are primarily intended for single-modality
systems. These standards can be inadequate for executing reproducible performance evaluations of
multimodal biometric systems such as those used in border control applications. Various configurations
are proposed for multimodal biometric systems, as described in ISO/IEC TR 24722. It is necessary to
clearly identify methods and requirements for multimodal biometric systems evaluation such as
variation of parameters and environmental factors that are to be described when reporting.
iv © ISO/IEC 2015 – All rights reserved

ISO/IEC 19795-2:2007/Amd.1:2015(E)
Information technology — Biometric performance testing
and reporting —
Part 2:
Testing methodologies for technology and scenario
evaluation
AMENDMENT 1: Testing of multimodal biometric
implementations
Add the following items to the list in the scope clause of ISO/IEC 19795-2:
— multimodal biometric specific requirements for technology evaluation and scenario evaluation;
— description of the structure and performance measures of multimodal biometric devices and systems;
— specification of biometric data collection and performance calculation methods;
— specification of reporting requirements.
Add the following item to the definitions clause in this part of ISO/IEC 19795:
4.4.2
multimodal FTE
MFTE
proportion of the population for whom the multimodal biometric system with a stated enrolment policy
fails to complete the enrolment process
EXAMPLE 1 For a policy which allows one of a set of modalities to be enrolled, MFTE would be the
proportion of subjects that fail to enrol in all the modalities. This results in lower effective FTE.
EXAMPLE 2 For a policy which requires enrolment in all considered modalities, MFTE would be the
proportion of subjects that fail to enrol in one or more modalities. This results in an effective FTE greater
than or equal to the higher of the individual modality FTEs.
ISO/IEC 19795-2:2007/Amd.1:2015(E)

Insert the following normative Annex into this part of ISO/IEC 19795:
Annex F
(normative)
Testing of multimodal biometric implementations
F.1   General
This annex specifies methods for evaluating and reporting the performance of multimodal biometric
algorithms and systems.
Multimodal biometric implementations might be used to meet the following objectives:
— to support users who cannot present one or more requested modalities to the system, in other
words, to improve failure-to-enrol rate;
— to improve biometric system throughput;
— to improve recognition performance (e.g. through reduction of false negative identification rates);
— to improve usability; and
— to increase robustness against presentation attacks.
ISO/IEC TR 24722 defines the following multimodal fusion levels:
— decision-level;
— score-level;
— feature-level;
— sample-level.
Multimodal fusion implementations differ across each level. Even when multimodal data are gathered
with identical sensors, results might differ based the fusion level implemented.
For this reason, the experimenter shall determine the system or application to be evaluated. An evaluation
shall clearly identify the fusion level implemented, the components of the multimodal implementation,
and requirements applicable to evaluations for each fusion level.
Two types of multimodal evaluations can be considered:
— evaluations in which the experimenter does require insight into the multimodal system;
— evaluations in which the experimenter does not require insight into the integrated multimodal system.
This annex focuses on evaluations in which the experimenter requires insight into the multimodal
system. F.1 and F.6 are applicable to all multimodal implementations. If the experimenter does not
require insight into the integrated multimodal system component shown in Figure F.2 or Figure F.4,
then the multimodal implementation can be tested without using this annex.
2 © ISO 2015 – All rights reserved

ISO/IEC 19795-2:2007/Amd.1:2015(E)

F.2     Fusion scheme identification information for repeatable evaluation
F.2.1   Decision-level fusion
F.2.1.1   General
An example of decision-level fusion is shown in Figure F.1. Decision-level fusion systems combine
decision results from separate biometric sub systems.
Comparison 1
Comparison 2
Figure F.1 — Decision-level fusion
NOTE Decision-level fusion systems might be used to improve false match rate (FMR) or false non-match rate
(FNMR).
Figure F.2 is an example of an integrated multimodal system with a decision output in which insight is
not provided into the implementation. This type of system can be tested without methods described
in this annex.
ISO/IEC 19795-2:2007/Amd.1:2015(E)

Figure F.2 — Fusion in an integrated multimodal system with decision output
F.2.1.2   Technology evaluation
Requirements for repeatability of decision-level fusion technology evaluation results are as follows:
— the decision fusion logic shall be identical;
— the function configurations (i.e. feature extraction, comparison, and decision) of Sub System 1 and
Sub System 2, respectively, shall remain consistent across all tests;
NOTE 1 Sub System 1 and Sub System 2 can have different function configurations, and user-specific thresholds
can differ for different users.
— the combination of Sample 1 and Sample 2 fed into each feature extraction function shall be identical;
— the combination of Template 1 and Template 2 shall be identical.
Consistent data selection methods for samples and templates are also required for evaluation
repeatability.
If Sub System 1 and Sub System 2 are independent and separate, the evaluation report should include
the following:
— identifying information for Sub System 1 and Sub System 2;
— identifying information for decision fusion logic;
— fusion level.
NOTE 2 See 6.4.2.
F.2.1.3   Scenario evaluation
Requirements for repeatability of decision-level fusion evaluation results are as follows:
— the function configurations (i.e. capture, feature extraction, comparison, and decision) of Sub
System 1 and Sub System 2, respectively, shall remain consistent across all tests;
4 © ISO 2015 – All rights reserved

ISO/IEC 19795-2:2007/Amd.1:2015(E)

NOTE 1 Sub System 1 and Sub System 2 can have different function configurations, and user-specific thresholds
may differ for different users.
— the decision fusion logic shall be identical;
— the combination of Sample 1 and Sample 2 fed into each feature extraction function shall be based
on the same subject and position (e.g. right iris);
— the combination of Template 1 and Template 2 shall be based on the same subject and position.
Consistent data selection methods for samples and templates are also required for evaluation
repeatability.
If Sub System 1 and Sub System 2 are independent and separate, the evaluation report should include
the following:
— identifying information for Sub System 1 and Sub System 2;
— identifying information for decision fusion function;
— fusion level.
NOTE 2 See 7.4.2.
F.2.2   Score-level fusion
F.2.2.1   General
Fusion on the score level is illustrated in Figure F.3. Score-level fusion systems utilize score results from
separate biometric subsystems.
CComparison 11
CComparison 22
Figure F.3 — Score-level fusion
Score-level fusion may use sample quality in scenario or technology evaluations.
NOTE 1 Score-level fusion systems might be used to improve false match rate (FMR) and false non-match rate
(FNMR).
Figure F.4 is an example of an integrated multimodal system with a score output in which insight is
not provided into the implementation. This type of system can be tested without methods described
in this annex.
ISO/IEC 19795-2:2007/Amd.1:2015(E)

Figure F.4 — Fusion in an integrated multimodal system with score output
F.2.2.2   Technology evaluation
Requirements for repeatability of score-level fusion technology evaluation results are as follows:
— the score fusion function and decision function shall be identical;
— the function configurations (i.e. feature extraction and comparison) of Sub System 1 and Sub System
2, respectively, shall remain consistent across all tests;
NOTE 1 Sub System 1 and Sub System 2 can have different function configurations, and user-specific thresholds
might differ for different users.
— the combination of Template 1 and Template 2 shall be identical;
— the combination of Sample 1 and Sample 2 fed into each feature extraction function shall be identical.
Requirements will be necessary for the data selection method for samples and templates, in order to
keep repeatability.
If Sub System 1 and Sub System 2 are independent and separate, the evaluation report should include
the following:
— identifying information for Sub System 1 and Sub System 2;
— identifying information for score fusion function and decision function;
— fusion level.
NOTE 2 See 6.4.2.
F.2.2.3 Scenario evaluation
Requirements for repeatability of score-level fusion scenario evaluation results can be stated as follows:
— the function configurations (i.e. capture, feature extraction, and comparison) of Sub System 1 and
Sub System 2, respectively, shall remain consistent across all tests;
NOTE 1 Sub System 1 and Sub System 2 can have different function configurations, and user-specific thresholds
can differ for different users.
— the score fusion function and decision function shall be identical;
— the combination of Template 1 and Template 2 shall be based on the same subject and position;
— the combination of Sample 1 and Sample 2 fed into each feature extraction function shall be based
on the same subject and position.
6 © ISO 2015 – All rights reserved

ISO/IEC 19795-2:2007/Amd.1:2015(E)

Consistent data selection methods for samples and templates are also required for evaluation
repeatability.
If Sub System 1 and Sub System 2 are independent and separate, the evaluation report should include
the following:
— identifying information for Sub System 1 and Sub System 2;
— identifying information for score fusion function and decision function;
— fusion level.
NOTE 2 See 7.4.2.
F.2.3   Feature-level fusion
F.2.3.1   General
Fusion on the feature level is illustrated in Figure F.5. Feature-level fusion systems utilize results from
separate feature extraction components.
Figure F.5 — Fusion on the feature level
F.2.3.2   Technology evaluation
Requirements for repeatability of feature-level fusion technology evaluation results are as follows:
— feature fusion function, comparison function, and decision function shall be identical;
— the function configurations (i.e. feature extraction) of Feature extraction 1 and Feature extraction
2, respectively, shall remain consistent across all tests;
NOTE 1 Feature extraction 1 and Feature extraction 2 can have different function configurations, and user-
specific thresholds can differ for different users.
— the combination of Sample 1 and Sample 2 fed into each feature extraction function shall be identical;
— the combination of Sample 1 and Sample 2 at the time of template creation shall be identical.
Consistent data selection methods for samples and templates are also required for evaluation
repeatability.
If Feature extraction 1 and Feature extraction 2 are independent and separate, the evaluation report
should include the following:
— identifying information for Feature extraction 1 and Feature extraction 2;
— identifying information for feature fusion function, comparison function and decision function;
— fusion level.
...