ANSI INCITS ISO IEC 29159-1-2010 Information technology Biometric calibration augmentation and fusion data Part 1 Fusion information format.pdf

1、 INCITS/ISO/IEC 29159-1-2010 Information technology Biometric calibration, augmentation and fusion data Part 1: Fusion information format INCITS/ISO/IEC 29159-1-2010 PDF disclaimer This PDF file may contain embedded typefaces. In accordance with Adobes licensing policy, this file may be printed or v

2、iewed 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 Adobes licensing policy. The ISO Central Secretariat accepts no liability

3、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 suitabl

4、e 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. Adopted by INCITS (InterNational Committee for Information Technology Standards) as an American National Standard. Date of ANSI Approval: 1/

5、4/2011 Published by American National Standards Institute, 25 West 43rd Street, New York, New York 10036 Copyright 2011 by Information Technology Industry Council (ITI). All rights reserved. These materials are subject to copyright claims of International Standardization Organization (ISO), Internat

6、ional Electrotechnical Commission (IEC), American National Standards Institute (ANSI), and Information Technology Industry Council (ITI). Not for resale. No part of this publication may be reproduced in any form, including an electronic retrieval system, without the prior written permission of ITI.

7、All requests pertaining to this standard should be submitted to ITI, 1250 Eye Street NW, Washington, DC 20005. Printed in the United States of America ii ITIC 2011 All rights reserved INCITS/ISO/IEC 29159-1-2010 ITIC 2011 All rights reserved iiiContents Page Foreword .v Introductionvi 1 Scope1 2 Con

8、formance .1 3 Normative references1 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 values4 6.4 Fusion header block4 7 Common elements 8 7.1 General .8 7.2 Parameter kind.8 7.3 Parameter origin 9 7.4 Dis

9、tributions 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 (in

10、formative) 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 Bibliography23 Figures Figure 1 Schematic representation of fusion information format usa

11、ge.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 INCITS/ISO/IEC 29159-1-2010 iv ITIC 2011 All rights reservedTables Table 1 Fusion information format record structure3 Table 2 Fusion header block s

12、tructure 3 Table 3 Type 1 record structure3 Table 4 Type 2 record structure4 Table 5 Type 3 record structure4 Table 6 Textual representation of numerical value 4 Table 7 The fusion header block 5 Table 8 CBEFF Product Identifiers.6 Table 9 Database identifiers6 Table 10 Database quality values.7 Tab

13、le 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

14、 Type 2 record format. 12 Table 20 Subtype C format. 14 Table 21 Type 3 record format. 14 Annex A (informative) Document Overview. 16 Table A.1 Fusion information format type taxonomy 16 Annex B (informative) Example Cumulative Distribution Functions 18 Annex C (informative) Use of pre-normalized da

15、ta. 20 Annex D (informative) Source for evaluation of spline. 22 INCITS/ISO/IEC 29159-1-2010 ITIC 2011 All rights reserved vForeword ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies (ISO member bodies). The work of preparing Internationa

16、l 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,

17、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 commi

18、ttees 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 po

19、ssibility 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.

20、ISO/IEC 29159 consists of the following parts, under the general title Information technology Biometric calibration, augmentation and fusion data: Fusion information format INCITS/ISO/IEC 29159-1-2010 vi ITIC 2011 All rights reservedIntroduction Biometric systems embed disparate technologies and com

21、parison 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 uniqu

22、ely 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 info

23、rmation 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

24、 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-presentat

25、ion (e.g. three images of a users 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 individuals biometric information to one or more enrolled sampl

26、es 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 score

27、s 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 com parison scores are combined directly to produce an output decision or score. In norm

28、alization-based processes, fusion is prec eded 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 en

29、tire 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 in

30、tegration 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)

31、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 sc

32、ore 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

33、 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 i

34、f 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, whet

35、her 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 INCITS/ISO/IEC 29159-1-2010 ITIC 2011

36、 All rights reserved viisystems, 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 interoperabilit

37、y 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

38、 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. Initialization Initialization Decision Logic Fusion module F

39、IF Record for Face Comparison Subsystem FIF Record for Finger Comparison Subsystem Biometric probe (image or template) generated for verification or identification Biometric reference (image or template) generated at enrolment Biometric probe (image or template) generated for verification or identif

40、ication Biometric reference (image or template) generated at enrolment Face comparison subsystem Finger comparison subsystem Score Score Component or data supplied by vendor CComponent or data supplied by vendor BComponent or data supplied by vendor A Figure 1 Schematic representation of fusion info

41、rmation format usage AMERICAN NATIONAL STANDARD INCITS/ISO/IEC 29159-1-2010 ITIC 2011 All rights reserved 1Information 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 t

42、hat 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

43、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 applic

44、ation 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 Biome

45、tric 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.

46、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. INCITS/ISO/IEC 29159-1-2010 2 ITIC 2011 All rights reserved4.2 cumulative distribution fu

47、nction 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

48、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

49、 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 con