ISO/IEC 29159-1:2010
(Main)Information technology — Biometric calibration, augmentation and fusion data — Part 1: Fusion information format
Information technology — Biometric calibration, augmentation and fusion data — Part 1: Fusion information format
ISO/IEC 29159-1:2010 specifies a biometric fusion information format that establishes machine readable data formats to describe the statistics of comparison score inputs to a fusion process. ISO/IEC 29159-1:2010 does not standardize comparison-score normalization processes, nor standardize or define fusion processes.
Technologies de l'information — Étalonnage biométrique, données d'augmentation et de fusion — Partie 1: Format d'information de fusion
General Information
Standards Content (Sample)
INTERNATIONAL ISO/IEC
STANDARD 29159-1
First edition
2010-09-01
Information technology — Biometric
calibration, augmentation and fusion
data —
Part 1:
Fusion information format
Technologies de l'information — Étalonnage biométrique, données
d'augmentation et de fusion —
Partie 1: Format d'information de fusion
Reference number
ISO/IEC 29159-1:2010(E)
©
ISO/IEC 2010
---------------------- Page: 1 ----------------------
ISO/IEC 29159-1:2010(E)
PDF disclaimer
This PDF file may contain embedded typefaces. In accordance with Adobe's licensing policy, this file may be printed or viewed but
shall not be edited unless the typefaces which are embedded are licensed to and installed on the computer performing the editing. In
downloading this file, parties accept therein the responsibility of not infringing Adobe's licensing policy. The ISO Central Secretariat
accepts no liability in this area.
Adobe is a trademark of Adobe Systems Incorporated.
Details of the software products used to create this PDF file can be found in the General Info relative to the file; the PDF-creation
parameters were optimized for printing. Every care has been taken to ensure that the file is suitable for use by ISO member bodies. In
the unlikely event that a problem relating to it is found, please inform the Central Secretariat at the address given below.
COPYRIGHT PROTECTED DOCUMENT
© ISO/IEC 2010
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means,
electronic or mechanical, including photocopying and microfilm, without permission in writing from either ISO at the address below or
ISO's member body in the country of the requester.
ISO copyright office
Case postale 56 • CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail copyright@iso.org
Web www.iso.org
Published in Switzerland
ii © ISO/IEC 2010 – All rights reserved
---------------------- Page: 2 ----------------------
ISO/IEC 29159-1:2010(E)
Contents Page
Foreword .v
Introduction.vi
1 Scope.1
2 Conformance .1
3 Normative references.1
4 Terms and definitions .1
5 Symbols and abbreviated terms .2
6 Fusion information format (FIF).3
6.1 Overview.3
6.2 Byte ordering .4
6.3 Numeric values .4
6.4 Fusion header block.4
7 Common elements .8
7.1 General .8
7.2 Parameter kind.8
7.3 Parameter origin .9
7.4 Distributions present .9
7.5 Number of comparisons .9
7.6 Pre-normalization flag.9
8 Type 1 record .10
8.1 Purpose .10
8.2 Format.10
8.3 Use case (Informative) .11
9 Type 2 record .12
9.1 Purpose .12
9.2 Format.12
9.3 Use case (Informative) .13
10 Type 3 record .13
10.1 Purpose .13
10.2 Format.14
Annex A (informative) Document Overview .16
Annex B (informative) Example Cumulative Distribution Functions.18
Annex C (informative) Use of pre-normalized data.20
Annex D (informative) Source for evaluation of spline .22
Bibliography.23
Figures
Figure 1 — Schematic representation of fusion information format usage.vii
Figure B.1 — Example CDFs and their spline representations .19
Figure C.1 — Example CDFs of internal comparison scores and pre-normalized scores .20
© ISO/IEC 2010 – All rights reserved iii
---------------------- Page: 3 ----------------------
ISO/IEC 29159-1:2010(E)
Tables
Table 1 — Fusion information format record structure.3
Table 2 — Fusion header block structure .3
Table 3 — Type 1 record structure.3
Table 4 — Type 2 record structure.4
Table 5 — Type 3 record structure.4
Table 6 — Textual representation of numerical value .4
Table 7 — The fusion header block .5
Table 8 — CBEFF Product Identifiers.6
Table 9 — Database identifiers.6
Table 10 — Database quality values.7
Table 11 — Score sense codes .8
Table 12 — Identifiers for statistical quantities .8
Table 13 — Origins of statistical data.9
Table 14 — Distribution information present.9
Table 15 — Pre-normalization codes.10
Table 16 — Subtype A format.10
Table 17 — Type 1 record format.10
Table 18 — Subtype B format.12
Table 19 — Type 2 record format.12
Table 20 — Subtype C format.14
Table 21 — Type 3 record format.14
Table A.1 — Fusion information format type taxonomy.16
iv © ISO/IEC 2010 – All rights reserved
---------------------- Page: 4 ----------------------
ISO/IEC 29159-1:2010(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.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.
The main task of technical committees is to prepare International Standards. Draft International Standards
adopted by the technical committees are circulated to the member bodies for voting. Publication as an
International Standard requires approval by at least 75 % of the member bodies casting a vote.
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.
ISO/IEC 29159-1 was prepared by Joint Technical Committee ISO/IEC JTC 1, Information technology,
Subcommittee SC 37, Biometrics.
ISO/IEC 29159 consists of the following parts, under the general title Information technology — Biometric
calibration, augmentation and fusion data:
⎯ Fusion information format
© ISO/IEC 2010 – All rights reserved v
---------------------- Page: 5 ----------------------
ISO/IEC 29159-1:2010(E)
Introduction
Biometric systems embed disparate technologies and comparison algorithms. Although some of these have
been published, most are entirely proprietary. Most current verification or identification applications employ a
single biometric modality. That is, information is acquired from a body part or an exhibited behavior with the
intent of more or less uniquely identifying the individual. For example, an access control system can image the
hand and use geometrical features. A social benefits program can collect fingerprints from applicants as input
to a one-to-many duplicate search. Different biometric modes offer varying amounts of discriminative
information and have different acquisition related problems. The effect is that biometric systems are to some
extent fallible and, moreover, they exhibit different failure modes. This affords opportunities to combine
technologies or algorithms to improve performance and/or usability. Such combination is known as fusion.
Fusion can be multi-modal (e.g. observing the biometric characteristics, face and finger), multi-algorithmic (e.g.
face recognition algorithms A and B), multi-instance (e.g. index finger and thumb), multi-sensorial (e.g. optical
and ultrasound fingerprint sensor) or multi-presentation (e.g. three images of a user's face).
This part of ISO/IEC 29159 addresses the most common and most readily implemented method of fusion:
score-level fusion. This is implemented after two or more systems have processed and matched an
individual's biometric information to one or more enrolled samples and produced scalar comparison scores as
output. The scores can be either genuine (same-person) or impostor (different-person) scores and a fusion
scheme is designed to combine such scores so that the class boundary between genuine and impostor scores
is refined.
Distributions of comparison scores are unique to each biometric comparison subsystem. Score ranges and
the shapes of the distributions can differ greatly. Fusion is often implemented in two ways.
⎯ In classification-based processes, the available comparison scores are combined directly to produce an
output decision or score.
⎯ In normalization-based processes, fusion is preceded by a transformation of each score to a common
domain. Simple normalization techniques based on statistical parameters such as the mean and standard
deviation are sometimes effective, but more sophisticated techniques utilize detailed knowledge of the
entire score distribution. The fusion information format (FIF) defined in this part of ISO/IEC 29159 is
intended to flexibly support any of the popular transformations. By establishing a standardized means of
data exchange, this part of ISO/IEC 29159 supports a modular approach to biometric systems integration
in which both the comparison and fusion algorithms remain protected as black-box pieces of intellectual
property. Thus this part of ISO/IEC 29159 envisages an application in which two (or more) underlying
acquisition and comparison technologies (hand geometry and fingerprint, for example) each generate a
score which is fed to a fusion module which has been initialized with an appropriate instance of the FIF
defined herein.
Figure 1 depicts the logical role of the records in a (notional) multimodal fusion process.
This part of ISO/IEC 29159 defines containers for the distributional score information from a comparison
subsystem. It does not allow for joint distributional data that can fully capture the statistical properties of
multivariate scores (i.e. those from two or more vendors' subsystems or modalities). This means that
multimodal fusion is not supported by a description of the joint distributions of the biometric scores. This is
often a minor limitation because different modalities are often assumed to be independent. Even when the
scores are not independent, as is the case for multi-algorithm applications, score-level fusion techniques often
remain effective, even if they are not optimal.
This part of ISO/IEC 29159 is intended to support interoperability and data interchange among biometrics
applications and systems. As such it specifies requirements that solve the complexities of applying biometrics
to a wide variety of personal recognition applications, whether such applications operate in an open systems
environment or consist of a single, closed system. Open systems are built on standards based, publicly
defined data formats, interfaces, and protocols to facilitate data interchange and interoperability with other
vi © ISO/IEC 2010 – All rights reserved
---------------------- Page: 6 ----------------------
ISO/IEC 29159-1:2010(E)
systems, which can include components of different design or manufacture. A closed system can also be built
on publicly defined standards, and can include components of different design or manufacture, but inherently
has no requirement for data interchange and interoperability with any other system.
Biometric data interchange format standards and biometric interface standards are both necessary to achieve
full data interchange and interoperability for biometric recognition in an open systems environment. The
biometric International Standards developed within JTC 1/SC 37 form a layered set of International Standards
consisting of biometric data interchange formats and biometric interfaces, as well as application profiles that
describe the use of these International Standards in specific application areas.
Decision
Logic
FIF Record for FIF Record for Finger
Fusion
Face Comparison Comparison
module
Subsystem Subsystem
Initialization Initialization
Biometric Biometric
Score Score
reference reference
(image or (image or
template) template)
generated at generated at
enrolment enrolment
Face Finger
comparison comparison
subsystem subsystem
Biometric
Biometric
probe (image
probe (image
or template)
or template)
generated for
generated for
verification or
verification or
identification
identification
Component or data Component or data Component or data
supplied by vendor A supplied by vendor C supplied by vendor B
Figure 1 — Schematic representation of fusion information format usage
© ISO/IEC 2010 – All rights reserved vii
---------------------- Page: 7 ----------------------
INTERNATIONAL STANDARD ISO/IEC 29159-1:2010(E)
Information technology — Biometric calibration, augmentation
and fusion data —
Part 1:
Fusion information format
1 Scope
This part of ISO/IEC 29159 specifies a biometric fusion information format that establishes machine readable
data formats to describe the statistics of comparison score inputs to a fusion process.
This part of ISO/IEC 29159 does not
⎯ standardize comparison-score normalization processes, nor
⎯ standardize or define fusion processes.
2 Conformance
Records are conformant to this part of ISO/IEC 29159 if they conform to all normative requirements of
Clause 6. This requires conformance to either Clause 8, 9, or 10, each of which requires conformance to the
stated subclauses of Clause 7.
3 Normative references
The following referenced documents are indispensable for the application of this document. For dated
references, only the edition cited applies. For undated references, the latest edition of the referenced
document (including any amendments) applies.
IEEE 754-2008, IEEE Standard for Floating-Point Arithmetic
ISO/IEC 19785-1:2006, Information technology — Common Biometric Exchange Formats Framework —
Part 1: Data element specification
ISO/IEC 19794-1:2006, Information technology — Biometric data interchange formats — Part 1: Framework
4 Terms and definitions
For the purposes of this document, the terms and definitions given in ISO/IEC 19794-1 and the following
apply.
4.1
biometric sample
analog or digital representation of biometric characteristics prior to biometric feature extraction
NOTE A biometric capture device is a biometric capture subsystem with a single component.
© ISO/IEC 2010 – All rights reserved 1
---------------------- Page: 8 ----------------------
ISO/IEC 29159-1:2010(E)
4.2
cumulative distribution function
probability that a variate takes on a value less than or equal to a number
4.3
genuine score
comparison score from comparison of two samples from one person
4.4
impostor score
comparison score from comparison of two samples from different persons
4.5
location parameter
generic measure of the position of a distribution
NOTE The location parameter is not necessarily the mean of a distribution.
4.6
probability density function
derivative of the cumulative distribution function
4.7
scale parameter
generic measure of the breadth of a distribution
NOTE The scale parameter is generally not the variance (nor the standard deviation) of a distribution.
4.8
comparison score
scalar output from biometric comparison subsystem
NOTE The term comparison score is used generically in this part of ISO/IEC 29159 for both distances (smaller
indicates greater likelihood that samples come from same person) and similarity scores (higher indicates same-person).
This distinction is conveyed explicitly in the score sense type defined in 6.4.9.
5 Symbols and abbreviated terms
For the purposes of this document, the following abbreviations apply. In tables that define binary record
structures the use of the symbol "M" in the status column indicates that the field itself is mandatory. A value of
"O" indicates the field is optional. This means that the bytes for the fields might not be present. In all cases
records can be parsed because the presence or absence of an optional field is recorded in a preceding field.
CBEFF Common Biometric Exchange Formats Framework
CDF cumulative distribution function
ECDF empirical cumulative distribution function
FAR false accept rate
FMR false match rate
FIF fusion information format
PDF probability density function
ID identifier
2 © ISO/IEC 2010 – All rights reserved
---------------------- Page: 9 ----------------------
ISO/IEC 29159-1:2010(E)
6 Fusion information format (FIF)
6.1 Overview
6.1.1 Record structure
The FIF record is used to support modularity in multi-modal biometric and decision support systems. Its format
is given in Table 1.
NOTE An application should establish a profile of part of the ISO/IEC 29159 standard. A default profile would
explicitly call out one of the record types.
Table 1 — Fusion information format record structure
Type 1 Record and/ Type 2 Record and/o Type 3 Record
Fusion header block
(see clause 6.1.3) or (see clause 6.1.4) r (see clause 6.1.5)
6.1.2 Header structure
The fusion header block structure defines the format of the record, and indicates the content. Its format is
given in Table 2.
Table 2 — Fusion header block structure
Fusion = Format Version Record Biometric
header Identifier Number Length Type
block
clause 6.4 6.4.2 6.4.3 6.4.4 6.4.5
25 bytes 4 4 4 3
Comparison Database Enrolment Verification Score Number of
Subsystem ID Database Database Sense Type
Product ID quality quality Instances
row continuation
6.4.6 6.4.7 6.4.8 6.4.8 6.4.9 6.4.10
4 2 1 1 1 1
6.1.3 Type 1 record structure
The fusion header block structure defines the format of the record, and indicates the content. Its format is
given in Table 3.
Table 3 — Type 1 record structure
Type 1 Record Type Distributions Impostor Distribution Genuine Distribution
=
Structure Present
Num Comp Loc Scale Num Comp Loc Scale
clause 8.2 8.2.3 8.2.4 8.2.5 8.2.6 8.2.7 8.2.8 8.2.9 8.2.10
26 or 50 bytes 1 1 4 10 10 4 10 10
6.1.4 Type 2 record structure
The fusion header block structure defines the format of the record, and indicates the content. Its format is
given in Table 4.
© ISO/IEC 2010 – All rights reserved 3
---------------------- Page: 10 ----------------------
ISO/IEC 29159-1:2010(E)
Table 4 — Type 2 record structure
Type 2 Record Structure Type Distributions Present Impostor Distribution CDF Genuine Distribution
=
CDF
clause 9.2 9.2.3 9.2.4 9.2.5 9.2.6
16N+13 or 32N+22 bytes 1 1 16N+11 16N+11
6.1.5 Type 3 record structure
The fusion header block structure defines the format of the record, and indicates the content. Its format is
given in Table 5.
Table 5 — Type 3 record structure
Type 3 Record Structure Type Distributions Present Impostor Distribution CDF Genuine Distribution
=
CDF
clause 10.2 10.2.3 10.2.4 10.2.5 10.2.6
16N-18 or 32N-38 bytes 1 1 16N-20 16N-20
6.2 Byte ordering
Within the FIF record, and all well-defined data blocks therein, all multi-byte quantities shall be stored in Big
Endian format. That is, the more significant bytes of any multi-byte quantity are stored at lower addresses in
memory than the less significant bytes.
EXAMPLE For example, the value 1025 (2 to the 10th power plus one) would be stored as first byte= 00000100b
and second byte=00000001b.
6.3 Numeric values
All numeric values present in the defined Types of this part of ISO/IEC 29159 are fixed-length unsigned
integer quantities, unless specified otherwise.
All numeric values given in the text of this part of ISO/IEC 29159 are decimals, unless preceded by 0x, to
indicate hexadecimal, or suffixed by a "b" to indicate binary.
Table 6 — Textual representation of numerical value
Example value Radix Decimal value
1010b 2 10
39 10 39
0xF5 16 245
Double precision numbers shall be conformant with IEEE 754.
NOTE The specification of IEEE 754 may not be sufficient to avoid numerical inaccuracy.
6.4 Fusion header block
6.4.1 General
The fusion header block of Table 7 shall be present as the first block in all FIF records.
4 © ISO/IEC 2010 – All rights reserved
---------------------- Page: 11 ----------------------
ISO/IEC 29159-1:2010(E)
Table 7 — The fusion header block
Field Status Size (bytes) Valid Values Notes
Format identifier M 4 0x46494600 ASCII string "FIF" with null terminator
Version number M 4 0x30313000 ASCII string "010" with null terminator
32
Record length M 4 1 ≤ L ≤ 2 -1 Length of entire record in bytes
Biometric type M 3 0 ≤ t ≤ 0x080000 The modality from which this record came from
Comparison subsystem
M 4 ≥ 0 Currently vendor specified ID.
product ID
Database ID M 2 ≥ 0 Currently vendor specified ID.
Enrollment sample M 1 [0-100],254,255 Aggregate qualities of the samples used
quality
to compute the comparison score statistics
Verification sample
M 1 [0-100],254,255
quality
Score sense M 1 0 or 1 Distances or similarities? See clause 6.4.9.
Number of type instances M 1 1 ≤ N ≤ 4 0 is not allowed
6.4.2 Format identifier
The (4-byte) format identifier shall begin with the three ASCII characters ‘FIF’ to identify the record as
following this part of ISO/IEC 29159, followed by a zero byte as a null string terminator.
6.4.3 Version number
The (4-byte) version number shall consist of three ASCII numerals followed by a zero byte as a NULL string
terminator. The value denotes the version of this part of ISO/IEC 29159.
The first and second character represents the major version number and the third character represents the
minor revision number.
The version number shall be 0x30313000 which is “010” – Version 1 revision 0.
6.4.4 Record length
The (4-byte) record length block shall be the combined length in bytes for the entire record. It shall be the sum
of the fusion header block length (25 bytes) and any Type 1, 2 or 3 record lengths that follow.
6.4.5 Biometric type
The (3-byte) value shall be taken from CBEFF's enumeration of biometric modalities given in
ISO/IEC 19785-1:2005, 6.5.6. This value allows an application to know what biometric modality is represented
by the instance of the fusion information format.
EXAMPLE For scores from an implementation using vein patterns this value would be 0x040000.
6.4.6 Comparison subsystem Product ID
The (4-bytes = 2 + 2 bytes) identifier of the comparison subsystem algorithm (i.e. comparing the biometric
probe with the biometric reference) that produced the scores information described in this FIF record shall be
recorded according to Table 8. These two values are the CBEFF Product Identifiers described in
ISO/IEC 19785-1:2006.
© ISO/IEC 2010 – All rights reserved 5
---------------------- Page: 12 ----------------------
ISO/IEC 29159-1:2010(E)
Table 8 — CBEFF Product Identifiers
Assigned by Length Description
Product ID (stored in the first two bytes) IBIA 2 bytes See ISO/IEC 19785-1:2006
Version Number (stored in the second two bytes) Vendor 2 bytes
NOTE 1 It can be argued that a Product ID field is not needed because the statistical information embodies all that is
needed by a fusion module. The field is however useful to support, for example, version control and caching.
NOTE 2 The Product ID alone is likely insufficient to execute score no
...
Questions, Comments and Discussion
Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.