ISO/IEC 30137-1:2019

INTERNATIONAL ISO/IEC
STANDARD 30137-1
First edition
2019-05
Information technology — Use of
biometrics in video surveillance
systems —
Part 1:
System design and specification
Technologies de l'information — Utilisation de la biométrie dans les
systèmes de vidéosurveillance —
Partie 1: Conception et spécification
Reference number
©
ISO/IEC 2019
© ISO/IEC 2019
All rights reserved. Unless otherwise specified, or required in the context of its implementation, 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
CP 401 • Ch. de Blandonnet 8
CH-1214 Vernier, Geneva
Phone: +41 22 749 01 11
Fax: +41 22 749 09 47
Email: copyright@iso.org
Website: www.iso.org
Published in Switzerland
ii © ISO/IEC 2019 – All rights reserved

Contents Page
Foreword .v
Introduction .vi
1 Scope . 1
2 Normative references . 2
3 Terms and definitions . 2
3.1 Target subject related terms . 2
3.2 VSS related terms . 2
3.3 Biometric system related terms . 4
3.4 Environment/scenario related terms . 4
3.5 Symbols and abbreviated terms. 5
4 Comparison of terms used in biometric systems with those used in video surveillance .5
5 Architecture . 6
6 Use cases . 7
6.1 General . 7
6.2 Post event use cases . 8
6.3 Real time use cases . 8
6.4 Enrolment use cases . 9
7 Specification of hardware and software . 9
7.1 General . 9
7.2 Physical environment . 9
7.3 Illumination environment .10
7.4 Inducing frontal view .10
7.5 Cameras and supporting infrastructure .10
7.5.1 Selection of cameras .10
7.5.2 Positioning of cameras . .11
7.5.3 Infrastructure considerations .16
7.6 Biometric software .17
7.6.1 General.17
7.6.2 Face detection software .17
7.6.3 Face comparison software .18
7.6.4 Algorithm selection and testing .18
7.6.5 Other (non-biometric) software .18
7.7 Computational requirements .18
7.7.1 General.18
7.7.2 Core biometric processes .19
7.7.3 Reducing computational expense .20
7.8 Specification for reference image database .20
7.8.1 General.20
7.8.2 Reference database size .20
7.8.3 Reference image quality .21
7.8.4 Reference database maintenance .21
8 Multiple camera operation .22
9 Interfaces to related software .22
10 Guidance for operator assistance .23
11 System design considerations .23
11.1 General .23
11.2 Establishing the business requirements .24
11.3 Site survey .24
11.4 Size and content of the watchlist .25
11.5 Performance requirements .26
© ISO/IEC 2019 – All rights reserved iii

11.5.1 General.26
11.5.2 Key metrics of performance .26
11.5.3 Presentation Attack Detection (PAD) performance metrics .27
11.6 Image data and metadata considerations .27
Annex A (informative) Other related (but non-biometric) video analytic techniques and
applications .28
Annex B (informative) Societal considerations and governance processes .31
Annex C (informative) Case study: The use of AFR with VSS for traveller triaging at the border .33
Annex D (informative) Video acquisition measurements .35
Bibliography .45
iv © ISO/IEC 2019 – All rights reserved

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.
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 document 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 and IEC 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) or the IEC
list of patent declarations received (see http: //patents .iec .ch).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation of the voluntary nature of standards, the meaning of ISO specific terms and
expressions related to conformity assessment, as well as information about ISO's adherence to the
World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT) see www .iso
.org/iso/foreword .html.
This document was prepared by Technical Committee ISO/IEC JTC 1, Information technology,
Subcommittee SC 37, Biometrics.
A list of all parts in the ISO 30137 series can be found on the ISO website.
Any feedback or questions on this document should be directed to the user’s national standards body. A
complete listing of these bodies can be found at www .iso .org/members .html.
© ISO/IEC 2019 – All rights reserved v

Introduction
Considerable improvements in the performance of automatic facial recognition (AFR) technologies
have resulted in applications such as automated border control using the facial images encoded in
e-passports and implemented in systems whereby the identity of a co-operative traveller is verified in
an environment designed for the collection of uniformly illuminated and optimally posed images. The
success of these first generation AFR systems has encouraged suppliers to consider other applications
where the environment for collection of images may be far from optimal. The inferior performance
in such less-controlled identification applications may necessitate a greater involvement by trained
personnel.
The ISO 30137 series provides guidance on the use of biometric technologies in video surveillance
systems (VSS), a framework for performance testing and reporting of such systems, and procedures for
establishing ground truth and annotating video data for testing purposes.
This document provides the architecture, use cases and system design. The use cases include real time
alerting to the presence of individuals of interest, law enforcement applications such as reviewing post-
event video footage from one or more cameras against pre-populated watchlists, commercial uses such
as the identification of individuals who are to be given preferential service, and faces added to (enrolled
in) a watchlist following observation of behaviours in the video material.
Other scenarios include measurement of crowd densities and determining numbers of individuals
traversing a given point. While these are not the focus of this document, they are closely related and
information on these is therefore included in Annex A.
vi © ISO/IEC 2019 – All rights reserved

INTERNATIONAL STANDARD ISO/IEC 30137-1:2019(E)
Information technology — Use of biometrics in video
surveillance systems —
Part 1:
System design and specification
1 Scope
The ISO 30137 series is applicable to the use of biometrics in VSS (also known as Closed Circuit Television
or CCTV systems) for a number of scenarios, including real-time operation against watchlists and in
post event analysis of video data. In most cases, the biometric mode of choice will be face recognition,
but this document also provides guidance for other modalities such as gait recognition.
This document:
— defines the key terms for use in the specification of biometric technologies in a VSS, including
metrics for defining performance;
— provides guidance on selection of camera types, placement of cameras, image specification etc. for
the operation of a biometric recognition capability in conjunction with a VSS;
— provides guidance on the composition of the gallery (or watchlist) against which facial images from
the VSS are compared, including the selection of appropriate images of sufficient quality, and the
size of the gallery in relation to performance requirements;
— makes recommendations on data formats for facial images and other relevant information (including
metadata) obtained from video footage, used in watchlist images, or from observations made by
human operators;
— establishes general principles for supporting the operator of the VSS, including user interfaces
and processes to ensure efficient and effective operation, and highlights the need to have suitably
trained personnel;
— highlights the need for robust governance processes to provide assurance that the implemented
security, privacy and personal data protection measures specific to the use of biometric technologies
with a VSS (e.g. internationally recognizable signage) are fit for purpose, and that societal
considerations are reflected in the deployed system.
This document also provides information on related recognition and detection tasks in a VSS such as:
— estimation of crowd densities;
— determining patterns of movement of individuals;
— identification of individuals appearing in more than one camera;
— use of other biometric modalities such as gait or iris;
— use of specialized software to infer attributes of individuals, e.g. estimation of gender and age;
— interfaces to other related functionality, e.g. video analytics to measure queue lengths or to alert for
abandoned baggage.
© ISO/IEC 2019 – All rights reserved 1

2 Normative references
There are no normative references in this document.
3 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
ISO and IEC maintain terminological databases for use in standardization at the following addresses:
— IEC Electropedia: available at http: //www .electropedia .org/
— ISO Online browsing platform: available at https: //www .iso .org/obp
3.1 Target subject related terms
3.1.1
operator
individual(s) responsible for day to day operation of the system
Note 1 to entry: This may include adjustment of the video surveillance cameras, selecting data suitable for use by
the biometric application, and acting on the output of the biometric comparison process.
3.1.2
presentation attack
presentation of an artefact or of human characteristics to a biometric capture subsystem in a fashion
that could interfere with the intended policy of the biometric system
3.1.3
target subject(s)
target(s)
individual(s) of interest
Note 1 to entry: A target subject will normally be someone already enrolled in a watchlist (3.1.4). However, this is
not always the case; in some scenarios they are a target because they are to be enrolled in a watchlist.
3.1.4
watchlist
list of individuals of interest (3.1.3) (and their associated reference images) for detection by the video
surveillance application
Note 1 to entry: The watchlist may be of individuals for whom an added service level is to be offered (e.g. VIPs or
premium customers). This is sometimes referred to as a “whitelist”.
Note 2 to entry: The watchlist may be a list of “wanted” individuals, e.g. individuals who should be denied access
to premises or services. This is sometimes referred to as a “blacklist”.
Note 3 to entry: A system may have multiple watchlists of different groups of target subjects, and with different
performance goals.
Note 4 to entry: In the case of target subject back-tracking (3.3.1) the watchlist will normally contain only one
target subject (3.1.3) (or in the case of a group of individuals of interest, a few target subjects).
3.2 VSS related terms
3.2.1
codec
computer program capable of encoding or decoding a digital data stream or signal
2 © ISO/IEC 2019 – All rights reserved

3.2.2
compression ratio
measure of the compressed file size to that of the uncompressed file size
3.2.3
dropped frames
frames from the video camera(s) that are not processed or are not available for facial detection and the
creation of templates
Note 1 to entry: Normally measured in terms of either the number of frames per second dropped, or the
percentage of the frames per second dropped.
3.2.4
frame
single image shown as part of a sequence of images in a video stream
3.2.5
frame rate
frequency (rate) at which an imaging device produces unique consecutive images called frames (3.2.4)
Note 1 to entry: Frame rate is normally expressed in frames per second (fps).
3.2.6
frame size
pixel dimensions of the frame described in terms of horizontal and vertical pixels, and which may also
be additionally described in terms of total megapixels
3.2.7
post-processing
steps performed after the biometric comparison process
EXAMPLE Triaging decisions based on fusion of quality and score metrics.
3.2.8
pre-processing
steps performed prior to the biometric comparison process
EXAMPLE Image quality enhancement, subject detection and feature extraction.
3.2.9
resolution
measure of the amount of detail that can be stored in an image
Note 1 to entry: Resolution is normally measured in pixels per millimetre.
3.2.10
subject tracking
process of aggregating multiple biometric samples for a single individual, possibly from multiple
cameras, to avoid producing separate detection alerts for the same target subject (3.1.3)
3.2.11
video management system
VMS
component of a video surveillance system (3.2.12) that collects video from cameras and other sources,
records that video to a storage device and provides an interface to both view the live video and to
randomly access recorded video according to time
3.2.12
video surveillance system
VSS
system consisting of camera equipment, monitoring and associated equipment for transmission and
controlling purposes, which may be necessary for the surveillance of a protected area
© ISO/IEC 2019 – All rights reserved 3

3.3 Biometric system related terms
3.3.1
back-tracking
act of finding the given image(s) of a face/individual by searching all video feeds where the individual
could have been seen
Note 1 to entry: Back-tracking may or may not use facial biometrics.
3.3.2
face detection
determination of the presence of faces within a video frame (3.2.4) and production of the location of
each face in the frame
Note 1 to entry: Face detection is the first step in the face recognition process.
3.3.3
post event analysis
non-realtime analysis of data previously captured by video surveillance cameras
EXAMPLE To identify possible suspects following an incident or event.
3.3.4
real time analysis
on-line processing of video surveillance data as it is captured
EXAMPLE To identify individuals held on a watchlist so that immediate action can be taken.
3.3.5
Wiegand
de-facto wiring standard commonly used to connect a card swipe mechanism to the rest of an electronic
entry system
3.3.6
zone of recognition
3-dimensional space within the field of view of the camera and in which the imaging conditions for
robust biometric recognition are met
Note 1 to entry: In general, the zone of recognition is smaller than the field of view of the camera, e.g. not all faces
in the field of view may be in focus and not every face in the field of view is imaged with the necessary inter-eye
distance (IED).
3.4 Environment/scenario related terms
3.4.1
attractor
visual or acoustic cue within the environment which encourages individuals to look in a particular
direction (i.e. towards the camera in a facial recognition application) in an attempt to improve
recognition performance
3.4.2
choke point
point of congestion or obstruction through which individuals pass
3.4.3
lux
measure of illumination intensity
4 © ISO/IEC 2019 – All rights reserved

3.5 Symbols and abbreviated terms
AFIS Automated Fingerprint Identification System
AFR Automated Facial Recognition
CCTV Closed Circuit Television (system), another term for video surveillance (system)
FPS Frames per Second
LFR Live Facial Recognition, real time automated facial recognition using video surveillance
cameras
GUI Graphical User Interface
HDR High Dynamic Range
IED Inter Eye Distance, the distance (usually measured in pixels) between the centres of the eyes
IP Internet Protocol
MTF Modulation Transfer Function
NIST National Institute of Standards and Technology
OSDP Open Supervised Device Protocol
PTZ Pan, Tilt and Zoom; a type of video surveillance camera that can be remotely adjusted (man-
ually by the operator or automatically by using dedicated software).
SFR Spatial Frequency Response
VMS Video Management System
VSS Video Surveillance System
4 Comparison of terms used in biometric systems with those used in video
surveillance
The video surveillance and biometrics communities both have well established vocabularies to describe
the various components of a system, but the same term may sometimes be interpreted differently.
While the terms listed above apply to this document, Table 1 below highlights some of those terms and
expressions where care needs to be taken when communicating with members of the video surveillance
community.
Table 1 — Comparison of terms used in biometric systems with those used in video surveillance
Term Definition within the context of automated Definition within the conventional use
biometric processing of human-led VSS, e.g. within the scope
of IEC 62676 series
Crowd monitor- For example, counting of individuals in a volume, The observation of a group to determine
ing or over a time interval collective behaviour or as part of a pro-
cess to detect anomalous activity
Detection and Biometric detection: the process of finding Target detection: the process of finding
localization instances of a particular biometric mode, while targets of interest, such as humans or
correctly rejecting all instances of imagery not cars, in a video feed
representing that biometric mode
Observation Tracking: the process of spatially locating a par- Target observation: the process of follow-
ticular biometric subject as it moves ing a particular target in a video feed
© ISO/IEC 2019 – All rights reserved 5

Table 1 (continued)
Term Definition within the context of automated Definition within the conventional use
biometric processing of human-led VSS, e.g. within the scope
of IEC 62676 series
Recognition The process for assigning a biometric identifier to The process of recognizing a familiar face,
a subject synonym for identification
Identification The process of determining a subject’s identity by The process of a human determining
comparing imagery of a biometric mode against a subject’s identity using available
a database formed from imagery of individuals. (printed) galleries, or use of identity cues
This generally includes not assigning an identi- (clothing)
fy when the target subject is not present in the
database
Verification The process of confirming a subject’s identity by The process of confirming a target’s
comparing imagery of a biometric mode against a identity
particular prior sample of a candidate individual
Inspection Human review of the output from an automat- Inspection: The detailed review of VSS
ed biometric system to assess an alert from the imagery to determine more detailed in-
biometric subsystem formation or characteristics, such as age
or sex of an individual, brand of clothing,
presence of jewellery
Alert An indication that an identifier for an enrolled An indication issued by a camera, opera-
subject has been returned by the biometric recog- tor or system that an event of interest has
nition process occurred
5 Architecture
Figure 1 — Components of a biometrically enabled VSS
6 © ISO/IEC 2019 – All rights reserved

Figure 1 shows the process flow in a typical biometrically enabled VSS with components such as:
1. Video surveillance cameras positioned to collect images in a form which supports comparison with
images on the watchlist.
2. A VMS and infrastructure to organise and transmit footage from a number of cameras to the main
server and storage system.
3. Software to detect and track faces (and/or other biometric features) in the video stream and to
create biometric feature sets in the format developed by the supplier of the biometric recognition
system. This can include feature sets created by combining features extracted from multiple face
images from a single individual, continuously updated as new video frames are processed.
4. Comparison and decision software, again likely to be proprietary to the supplier of the biometric
system, which determines whether the system has recognised an individual on the watchlist. The
match criteria and decision thresholds may be different for groups of individuals on the watchlist;
e.g. some may be considered low risk, with only minimal implications should they not be recognised
by the system, whereas for others it may be imperative that they are recognised as soon as possible.
5. Alerts generated by the automated system are passed to the human operator for assessment.
6. An operator support environment to aid in making decisions on whether an alert should be followed
up (and how) or rejected as a false alert.
7. Links to analytics systems to record the event and decisions taken, and to provide access to other
information which may assist in disposal of the instance of recognition, e.g. previous instances of a
similar match to the individual on the watchlist, and guidance on the appropriate action to be taken.
8. A systems management “bus” which enables configuration and operation of the key components in
the biometric recognition system according to threat level, workload of human operators, time of
day, etc. and supports the merging of recognitions between cameras across the surveillance domain.
Figure 1 shows an example of a server-centric architecture. However, there are other models available,
such as distributed architectures using edge computing (where part of the processing is done in the
video camera of the VSS) or where cameras and computing resources are available within smart devices
such as smartphones and PCs.
6 Use cases
6.1 General
This section provides examples of some of the different ways in which biometrics can be used in
conjunction with VSS to support business needs across a range of organisations, including:
— police and law enforcement (and private security companies, such as those operating shopping
malls and car parks) to alert to the presence of individuals of interest;
— police and law enforcement to manage the identification of individuals in video surveillance footage
collected after a notable event or incident;
— commercial usage to alert to the presence of individuals of interest for whom special or differentiated
levels of service are to be provided;
— commercial or government systems to manage the flow of individuals or queues, e.g. in accordance
with agreed service levels;
— border services and client support organizations for quality assurance and customer support, e.g.
following a complaint or an incident.
The use cases can be broken down into three broad categories, namely ‘post event’, ‘real-time’ and
‘enrolment’ applications (enrolment may be real time or post event). The following sections provide
© ISO/IEC 2019 – All rights reserved 7

examples of some common use cases, described in terms of performance objectives and the roles played
by various components of the system, including the responsibilities of the system operator.
6.2 Post event use cases
In post event use cases the performance objective is the reliable detection, automated feature extraction
and searching of large numbers of target subjects against one or more watchlists or databases in an
attempt to identify possible suspects, with a high probability that the candidate list returned by the
biometric subsystem includes (at a high rank) those target subjects that have a matching template
stored in the watchlist.
These use cases are challenging because in many cases the quality and positioning of the video cameras
will be beyond the control of the operator of the biometric subsystem, and they will not have been
installed with biometric applications in mind.
The operator normally has an “expert” role within the end to end process, selecting images suitable
for submission as probes and examining candidates returned following a search of the database. They
may be trained in facial comparison techniques, and the decision-making process may be supported by
dedicated image analysis tools. In cases such as backtracking or clustering (linking images of the same
subjects together) the operator may also make use of other visual information (e.g. the individual’s
clothes and relative location of cameras) to help them to confirm or refute potential matches.
Examples of post event use cases include:
— post event analysis of recorded video surveillance material (from one or more cameras) processed
with the use of biometric recognition software to identify one or more individuals in frames or
sequences (using one or more reference images);
— post event analysis of recorded video surveillance material from more than one camera in which
an individual (whether identified or not) is tracked (either forwards or backwards in time) and
between cameras. This may involve more than just biometric applications, for example video
analytics software;
— retrospective clustering — detecting and extracting faces from multiple sources of video for the
purposes of clustering imagery sources of the same individual(s) together. This will normally need
to be an automated process due to large numbers of subjects appearing in multiple video streams,
although a human operator may subsequently review the results and intervene where they find
subjects who have been wrongly classified.
6.3 Real time use cases
In real time applications the performance objective is a high probability of the system alerting for
target subjects with a matching template in a watchlist, and a low probability of an alert for subjects
not in the watchlist. The watchlist will typically consist of a subset of images drawn from a larger image
database, and that have been chosen to address a specific business objective.
These use cases are challenging because of the large amount of data that needs to be processed,
especially if the system involves multiple cameras with multiple subjects in each frame. Not only does
this present a challenge in terms of search accuracy, it is vital that the end to end response time is fast
enough to enable effective action to be taken when an alert is generated.
The role of the operator will typically be to assess any alerts from the biometric subsystem and to make
an initial decision as to whether the alert is genuine or if it is a false match. They will usually also be
responsible for instigating further action as appropriate, such as directing resources on the ground to
detain or speak to the target subject in order to formally confirm their identity.
Examples of real time use cases include:
— alerting in real time (or near real time) to the presence of an individual traversing the field of view
of a video surveillance camera, identified by the biometric subsystem as being someone whose
8 © ISO/IEC 2019 – All rights reserved

biometric data (e.g. a facial image) has previously been stored as a reference in a watchlist. For
example, checking individuals entering a building or disembarking a plane or train against a
watchlist, with the aim of either bestowing or denying particular privileges, and monitoring of
video surveillance by law enforcement agencies for the purposes of crime prevention and public
safety, This use case is sometimes referred to as Live Facial Recognition (LFR). A practical example
illustrating the use of LFR can be found in Annex C;
— real time tracking of a particular individual of interest between the fields of view of a number of
cameras, some of which may not overlap.
6.4 Enrolment use cases
In these use cases the goal is the successful enrolment of target subjects of interest into a database
or watchlist, such that the biometric templates created are of sufficient quality for the intended use.
Prior to enrolment, a biometric search may be carried out to determine if the target subject is already
enrolled.
The operator may have a role in selecting the best quality images for enrolment, but in many cases
the process will be fully automated. Machine learning and cognitive computing can be applied to help
ensure that the best available images are selected for enrolment.
Examples of enrolment use cases include:
— enrolment (into a watchlist) of individuals who enter a protected zone or repeatedly visit the
same area;
— “time clocking” individuals in situations where there is an interest in knowing how long they spend
in a particular area, for example to monitor the time of service or queue length;
— enrolment of individuals traversing the field of view of a video surveillance camera into a database
in order to support a watchlist application for future use within the same system, or to use in
conjunction with other biometric applications and databases.
7 Specification of hardware and software
7.1 General
The IEC 62676 series already provides extensive information on camera selection, positioning, network
bandwidth, performance considerations, storage requirements etc. for traditional (i.e. non-biometric)
applications. This document therefore focuses on those aspects of the hardware and software
components of the VSS that have a direct bearing on the performance of the biometric subsystem.
It is important to note that what may be an ideal set up for a conventional VSS may produce images that
are very poorly suited for use in a biometric application. While the following recommendations are
primarily applicable to an AFR system, they can also be adapted for other modes.
7.2 Physical environment
In many cases the environment in which the VSS is intended to operate will be beyond the control of
those responsible for deploying or operating the cameras. Careful positioning of the cameras may help
(see 7.5.2), but where it is also possible to exert some influence over the environment where the system
operates, the following points should be considered:
— uneven floors and steps should generally be avoided as changes of angle/height often cause
individuals to look down, thus making it hard to obtain usable images of the face;
— barriers may be introduced to modify the flow of individuals through the environment, ensuring
they all pass through the field of view of the camera(s) at the correct distance and moving towards
(in the case of an AFR system) the camera. Such techniques, together with careful positioning of the
© ISO/IEC 2019 – All rights reserved 9

cameras can increase the amount of time an individual is within the field of view which will in turn
improve the performance of the system;
— choke points may be introduced to reduce the number of individuals passing through the field of
view of the camera at any one time, thereby reducing the number of target subjects that need to be
processed simultaneously by the biometric application. Choke points can also improve biometric
sample collecting by limiting the speed with which target subjects move through a capture area, by
improving the lighting at that location, and creating a situation where the pose of the target subject
to the camera(s) is more favourable.
The introduction of barriers or choke points may have negative implications for individuals moving
through the environment. Due consideration should be given to the need for usability, accessibility and
user friendliness, and the balance between these factors and the need to obtain high quality images in
order to maximise system performance should be determined on a case by case basis. See Annex B for
more information on societal aspects to be considered when employing such techniques.
7.3 Illumination environment
Sufficient illumination is needed to support biometric processing. When possible, the following points
should be considered:
— areas near windows or in sunlight should generally be avoided as the lighting cannot be controlled
and will vary with the time of the day/year and prevailing weather conditions. Shaded or artificially
illuminated areas generally produce better results;
— additional lighting may be introduced to raise overall light levels and also to ensure balanced
illumination across the faces, with no strong shadows or excessively bright highlights; additional
lighting also allows faster shutter speeds to be used, helping to avoid motion blur in the video;
— near-infrared lighting may also be used (in conjunction with surveillance cameras that can detect
those wavelengths) to help reduce shadows and to improve biometric sample collection under low
light conditions.
7.4 Inducing frontal view
AFR is highly sensitive to the orientation of the head relative to the optical axis of the camera. It is
often possible to modify subject behaviour by installing “attractors” that encourage people to look in a
particular direction, e.g. upwards or towards specific camera position(s).
When possible, the environment should also be inspected to determine if any sources of distraction
exist, e.g. for example an extraneous television screen that could undermine the biometric capture
process by adversely taking attention from the intended direction of view.
7.5 Cameras and supporting infrastructure
7.5.1 Selection of cameras
Camera and lens combinations should be selected such that the image resolution, frame rate, field of
view and low-level light performance are capable of providing images of sufficient quality for use in the
intended biometric application.
Several different quantifiable metrics are necessary to assess the performance of a video camera and
associated system producing the video stream for video surveillance.
The spatial resolution of the video camera is one of the most important factors in determining the
quality of an image captured by a VSS. A measure of spatial resolution is the MTF. The camera’s original
image‘s MTF20 should be at least 0,4 cycles per pixel. The original image is the same as described in
Annex D and refers to the unencoded signal.
10 © ISO/IEC 2019 – All rights reserved

In
...