ISO/IEC 19794-2:2005/Amd 1:2010

INTERNATIONAL ISO/IEC
STANDARD 19794-2
First edition
2005-09-15
AMENDMENT 1
2010-04-01

Information technology — Biometric data
interchange formats —
Part 2:
Finger minutiae data
AMENDMENT 1: Detailed description of
finger minutiae location, direction, and type
Technologies de l'information — Formats d'échange de données
biométriques —
Partie 2: Données du point caractéristique du doigt
AMENDEMENT 1: Description détaillée du point caractéristique du
doigt, direction et type




Reference number
ISO/IEC 19794-2:2005/Amd.1:2010(E)
©
ISO/IEC 2010

---------------------- Page: 1 ----------------------
ISO/IEC 19794-2:2005/Amd.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 19794-2:2005/Amd.1:2010(E)
Foreword
ISO (the International Organization for Standardization) and IEC (the International Electrotechnical
Commission) form the specialized system for worldwide standardization. National bodies that are members of
ISO or IEC participate in the development of International Standards through technical committees
established by the respective organization to deal with particular fields of technical activity. ISO and IEC
technical committees collaborate in fields of mutual interest. Other international organizations, governmental
and non-governmental, in liaison with ISO and IEC, also take part in the work. In the field of information
technology, ISO and IEC have established a joint technical committee, ISO/IEC JTC 1.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.
The main task of the joint technical committee is to prepare International Standards. Draft International
Standards adopted by the joint technical committee are circulated to national bodies for voting. Publication as
an International Standard requires approval by at least 75 % of the national 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 and IEC shall not be held responsible for identifying any or all such patent rights.
Amendment 1 to ISO/IEC 19794-2:2005 was prepared by Joint Technical Committee ISO/IEC JTC 1,
Information technology, Subcommittee SC 37, Biometrics.

© ISO/IEC 2010 – All rights reserved iii

---------------------- Page: 3 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)

Information technology — Biometric data interchange
formats —
Part 2:
Finger minutiae data
AMENDMENT 1: Detailed description of finger minutiae location,
direction, and type
IMPORTANT — The electronic file of this document contains colours which are considered to be
useful for the correct understanding of the document. Users should therefore consider printing this
document using a colour printer.
Page 39
Insert the following new annex after D.3:
Annex E
(informative)

Detailed description of finger minutiae location, direction, and type
E.1 Scope
Even if all conform to this part of ISO/IEC 19794, different minutiae data blocks extracted from the same finger
image may differ not only in the exact locations, the directions, and the types of those minutiae that they have
in common, but also in the number of minutiae they contain, especially in blurred fingerprint regions where
even the "manual" detection of minutiae is hard. The description of the minutia location in 6.4 refers to a
single-pixel-wide skeleton of the friction ridges. The minutia direction is defined in 6.4, based on tangents to
the skeleton. The skeletonisation algorithm itself is not described and also the method to determine the
tangents is left open.
The scope of this informative annex is to provide a more precise definition of location, direction, and type of
minutiae in gray-scale finger images and a detailed description of the quality field. It enhances the readability
of this part of ISO/IEC 19794 and decreases the possibility of misinterpretation. The standardisation of
algorithms is out of scope of this informative annex. This informative annex does not supersede the existing
standard.
© ISO/IEC 2010 – All rights reserved 1

---------------------- Page: 4 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)
E.2 Terms and definitions
For the purposes of this informative annex, the following terms and definitions apply.
E.2.1
4-neighbour of a pixel p
pixel that is the top, bottom, left, or right neighbour of p
EXAMPLE The pixels e, f, g, and h in Figure E.1 are 4-neighbours of pixel p.
a e
b
h p
f
d g c

Figure E.1 — 4- and 8-neighbours of a pixel p
E.2.2
4-path from pixel p to pixel p
0 n
sequence of pixels (p , p , p , …, p ) such that p is a 4-neighbour of p
0 1 2 n i i-1
E.2.3
4-connected set of pixels
set S of pixels such that for any two pixels p, q ∈ S there exists a 4-path from p to q
E.2.4
8-neighbour of a pixel p
pixel that is a 4-neighbour or a diagonal (top-left, top-right, bottom-left, or bottom-right) neighbour of p
EXAMPLE The pixels a, b, c, d, e, f, g, and h in Figure E.1 are 8-neighbours of pixel p.
E.2.5
8-path from pixel p to pixel p
0 n
sequence of pixels (p , p , p , …, p ) such that p is an 8-neighbour of p
0 1 2 n i i-1
E.2.6
8-connected set of pixels
set S of pixels such that for any two pixels p, q ∈ S there exists an 8-path from p to q
E.2.7
border ∂S of a set of pixels S
subset ∂S ={}x ∈ S : x is 4-neighbour of q, q ∉ S of pixels of S that are 4-neighbours of pixels outside S
E.3 Minutiae detection strategy
E.3.1 "Liberal-conservative" spectrum
Minutia detection algorithms may use different discriminative practices in the minutia detection strategy. A
liberal minutia detection strategy is supposed to detect a large number of minutiae which will increase the
probability to include spurious minutiae while a conservative strategy will detect only a few minutiae and
increase the probability to miss some. The following subclauses provide an explanation of some types of
spurious (false) minutiae which may result from the use of a ‘liberal’ strategy but which may not be detected if
a more ‘conservative’ strategy is employed.
2 © ISO/IEC 2010 – All rights reserved

---------------------- Page: 5 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)
The following images show examples of applying a conservative or liberal minutia detection strategy to the
same sample images. These examples are not meant to suggest a liberal or conservative strategy. The best
detection strategy for a particular application depends on the business processes and their associated
security requirements that the biometric components of the system are designed to support or enable.

Figure E.2 — Liberal minutia detection (left) versus conservative minutia detection (right)


Figure E.3 — Liberal minutia detection (left) versus conservative minutia detection (right)
© ISO/IEC 2010 – All rights reserved 3

---------------------- Page: 6 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)

Figure E.4 — Liberal minutia detection (left) versus conservative minutia detection (right)


Figure E.5 — Liberal minutia detection (left) versus conservative minutia detection (right)
4 © ISO/IEC 2010 – All rights reserved

---------------------- Page: 7 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)

Figure E.6 — Liberal minutia detection (left) versus conservative minutia detection (right)
E.3.2 Fingerprint boundary
No minutia should be set outside the fingerprint boundary.
Minutiae may be set below the first phalange, even it is not the usual case.
E.3.3 Sweat pore
No minutia should be set at a sweat pore. A pore could happen to lie at the position of the forking of a friction
ridge (bifurcation, see Figure E.16 below), but a sweat pore without connectivity to three legs must not be
misinterpreted as a minutia.
E.3.4 Touching ridges
No minutia should be set where thick ridges touch each other.
E.3.5 Incipient ridge
No minutia should be set at an incipient (very short and thin) ridge.
E.3.6 Crease
No minutia should be set at a crease (accidental interruption of ridges).
E.3.7 Core
No minutia should be set at a core.
A core represents a singularity in the direction field, hence a proper angle value cannot be assigned to this
location.
NOTE Information about cores can be expressed in a standardised way in the extended data block (see 8.5.3).
© ISO/IEC 2010 – All rights reserved 5

---------------------- Page: 8 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)
E.3.8 Delta
No minutia should be set at a delta.
A delta represents a singularity in the direction field, hence a proper angle value cannot be assigned to this
location.
NOTE Information about deltas can be expressed in a standardised way in the extended data block (see 8.5.3).
E.4 Minutia characteristics
E.4.1 Rationale
This document shall not standardize certain algorithms as laid down in the scope. The guidelines to find the
best minutia position and location require some methodology in description. Examples of two independent
methods for determining the location and orientation of minutiae are presented in this document. The first is
commonly known as the ridge gradient method while the second is referred to as the valley skeletal bifurcation
method, which is popular in the AFIS industry. Without loss of generality, the ridge gradient method will focus
on ridge ends and ridge bifurcations and the valley skeletal bifurcation method will describe valley bifurcations
and ridge bifurcations in this document, i.e. the choice of the method finally depends on the specific format
type to be used.
E.4.2 Minutia type
The minutia type cannot be determined reliably in some occasions.
EXAMPLE Due to varying contact pressure while acquiring the fingerprint and due to different image binarisation
approaches, a ridge ending may join an adjacent ridge, giving the impression of a ridge bifurcation.
The minutiae type “other” should only be used if neither of the other two minutiae types, “ridge ending” and
“ridge bifurcation”, can reliably be assigned to a minutia.
E.4.3 Minutia location tools
E.4.3.1 Consideration of the resolution of the underlying finger image
For the minutiae location, a correct handling of the resolution of the underlying finger image is important. The
minutiae extraction algorithm should be able to determine the resolution of the underlying finger image in a
reliable way (e.g. from a fingerprint-sensor configuration file). For minutiae data in the finger minutiae record
format, this resolution shall be stored in the X and Y resolution fields within the record header. For minutiae
data in the on-card-biometric-comparison format, the resolution of the underlying finger image shall be used
when calculating the X and Y coordinates of the location in the prescribed metric dimension units out of their
pixel values. For conversion between format types, the resolution shall be taken from, or stored in, the X and
Y resolution fields within the record header.
E.4.3.2 Image binarisation
Every gray scale fingerprint image can be transformed into a binary image. This is common practice in image
processing. Every pixel is assigned black if its gray scale value is darker than a threshold (such as the
average gray scale value) and white if its gray scale value is lighter than the threshold. Most professional
finger image processing implementations use sophisticated methods such as location-dependent thresholds
to come to a binary image. A binary image separates the image pixels into two categories: ridges and valleys.
Without loss of generality, black pixels refer to ridges in the following text.
6 © ISO/IEC 2010 – All rights reserved

---------------------- Page: 9 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)

captured raw image binary image (truncated)
Figure E.7 — Raw image vs. binary image
E.4.3.3 Image skeletonization
Skeletonization is a standard procedure in graphing practice. It produces a single-pixel-wide skeleton from a
binary image. Several skeletonization methods are reported in literature. The process yields either a
4-connected or 8-connected skeleton. Figure E.8 shows a sample image and its skeleton.

Figure E.8 — Binary image and ridge skeleton, from [14]
© ISO/IEC 2010 – All rights reserved 7

---------------------- Page: 10 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)
As valley skeletons will also be used in this standard, Figure E.9 depicts the valley skeleton of the same
image.

Figure E.9 — Binary image and valley skeleton, from [14]
E.4.3.4 Ridge Flow Direction
Every fingerprint image has a well defined directional image expressing the local dominant ridge flow direction.
Methods to compute a directional image are reported in literature, e.g. [12]. Figure E.10 shows a fingerprint
image and its directional image.

Figure E.10 — Raw image and pixel-wise directional image, from [14]
Most current fingerprint minutiae detection algorithms are working with a block-wise directional image rather
than a pixel-wise. The original directional image is therefore divided into blocks and the most common
orientation within a block becomes the orientation for the whole block. Figure E.11 shows the block directional
image.
8 © ISO/IEC 2010 – All rights reserved

---------------------- Page: 11 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)


Figure E.11 — Raw image and block-wise directional image, from [14]
E.4.4 Ridge gradient method
The ridge gradient method relies on moving along the ridge line until a minutia condition occurs, which is
either forking or ending of the ridge. It was originally reported for gray-scale images [13], but is described here
for binary images to simplify the procedure, which has only a descriptive nature in this part of ISO/IEC 19794.
E.4.4.1 Minutia location at a ridge skeleton endpoint
Friction ridges in a binary image have a well defined border: Black pixels with at least one white pixel as
4-neighbour are border pixels of a friction ridge. This ensures that the border is at least 8-connected. The
border of a ridge skeleton endpoint is depicted in Figure E.12.
© ISO/IEC 2010 – All rights reserved 9

---------------------- Page: 12 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)

Figure E.12 — Border of a ridge skeleton endpoint
Skeletonisation algorithms should reduce the friction ridges of a binary fingerprint image to a single pixel wide
skeleton. Instead of discussing how the skeletonisation algorithm should work, another approach to describing
where to place the minutiae is preferred. Every ridge has a dominant ridge flow direction. The dominant ridge
flow direction (see E.4.3.4) line meets the border of the ridge in a single pixel of the binary image. This pixel is
considered the optimal minutia location. See Figure E.13.

Figure E.13 — Minutia location on a ridge skeleton endpoint
E.4.4.2 Minutia location at a ridge skeleton bifurcation point
The border at a ridge bifurcation is well defined. The three parts of the border are not necessarily connected.
The border of a ridge skeleton bifurcation point is depicted in Figure E.14.
10 © ISO/IEC 2010 – All rights reserved

---------------------- Page: 13 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)

Figure E.14 — Border of a ridge skeleton bifurcation point
The dominant ridge flow direction lines along all three legs meet in a single point or form a small triangle within
the ridge. The single intersecting point is considered the optimal minutia location. It is most likely half a ridge
width from the border pixel where an acute angle is enclosed by neighbouring legs. It could happen rarely and
depending on the image resolution that the three legs do not intersect in a single pixel but form a small
triangle at intersection. In this case, the optimal minutia position is the center of mass of the triangle. See
Figure E.15.

Figure E.15 — Minutia location on a ridge skeleton bifurcation point
E.4.4.3 Minutia location at a valley skeleton bifurcation point
Analogical rules as for ridge skeleton bifurcation points apply also for valley skeleton bifurcation points. The
valley skeleton birfurcation point is described in detail in the section using the valley skeletal bifurcation
method to determine minutia position and orientation.
© ISO/IEC 2010 – All rights reserved 11

---------------------- Page: 14 ----------------------
ISO/IEC 19794-2:2005/Amd.1:2010(E)
E.4.4.4 Minutia direction at a ridge skeleton endpoint
This part
...