GB/T 31488-2015 (GB/T31488-2015, GBT 31488-2015, GBT31488-2015)
| Standard ID | Contents [version] | USD | STEP2 | [PDF] delivered in | Standard Title (Description) | Status | PDF |
| GB/T 31488-2015 | English | 135 |
Add to Cart
|
0--9 seconds. Auto-delivery
|
Technical requirements for face identification of video surveillance in security systems
| Valid |
GB/T 31488-2015
|
Standards related to: GB/T 31488-2015
| Standard ID | GB/T 31488-2015 (GB/T31488-2015) | | Description (Translated English) | Technical requirements for face identification of video surveillance in security systems | | Sector / Industry | National Standard (Recommended) | | Classification of Chinese Standard | A91 | | Classification of International Standard | 13.310 | | Word Count Estimation | 14,190 | | Date of Issue | 2015-05-15 | | Date of Implementation | 2015-12-01 | | Quoted Standard | GB/T 25724-2010; GA/T 893; GA/T 922.2-2011; ITU-T H.264 | | Drafting Organization | Beijing in the security shield technology development company; First Research Institute of Ministry of Public Security; Institute of Automation, Chinese Academy of Sciences; Beijing Hisign tech Co., Ltd.; Institute of Computing Technology; Beijing Digital Technology Co., Ltd. Olsen | | Administrative Organization | National Security and Protection Alarm System Standardization Technical Committee biometric identification applications Technical Committee ���� (SAC/TC 100/SC 2) | | Regulation (derived from) | National Standard Announcement 2015 No. 15 | | Proposing organization | Ministry of public security | | Issuing agency(ies) | Administration of Quality Supervision, Inspection and Quarantine of People's Republic of China; Standardization Administration of China | | Summary | This standard specifies the basic structure of security video surveillance face recognition system, functional requirements, performance requirements and test methods. This standard applies to design security for the purpose of face recognition video surveillance systems, project acceptance and related product development. Other areas of video surveillance face recognition system can use and reference. | GB/T 31488-2015
GB
NATIONAL STANDARD OF THE
PEOPLE’S REPUBLIC OF CHINA
ICS 13.310
A 91
Technical Requirements for Face Identification of
Video Surveillance in Security Systems
ISSUED ON: MAY 15, 2015
IMPLEMENTED ON: DECEMBER 1, 2015
Issued by: General Administration of Quality Supervision, Inspection and
Quarantine;
Standardization Administration of the People’s Republic of
China.
Table of Contents
Foreword ... 3
1 Scope ... 4
2 Normative References ... 4
3 Terms and Definitions ... 4
4 Basic Structure of System ... 6
5 Functional Requirements of System ... 7
5.1 Face Registration ... 7
5.2 Face Image Acquisition ... 7
5.3 Face Comparison ... 8
5.4 Result Analysis ... 8
5.5 Alert Output ... 8
5.6 Watch List Management ... 8
5.7 Alert Record Management ... 8
5.8 Parameter Settings ... 8
5.9 User Management ... 9
5.10 System Log ... 9
6 Performance Requirements of System ... 9
7 Test Methods of System ... 10
7.1 Design Principle of Test... 10
7.2 Functional Test Methods ... 10
7.3 Performance Test Methods ... 13
7.4 Test Report ... 16
Appendix A (informative) Template of Test Report of Video Surveillance Face
Recognition System ... 17
Bibliography ... 19
Technical Requirements for Face Identification of
Video Surveillance in Security Systems
1 Scope
This Standard specifies the basic composition, functional requirements, performance
requirements and test methods of face recognition of video surveillance in security
systems.
This Standard is applicable to the scheme design, project acceptance inspection and
relevant product development of face recognition of video surveillance for the purpose
of security precautions. The face recognition of video surveillance in other fields may
also take this as a reference.
2 Normative References
The following documents are indispensable to the application of this document. In
terms of references with a specified date, only versions with a specified date are
applicable to this document. In terms of references without a specified date, the latest
version (including all the modifications) is applicable to this document.
GB/T 25724-2010 Technical Specification of Surveillance Video and Audio Coding
GA/T 893 Terms for Biometric Recognition in Security and Protection Systems
GA/T 922.2-2011 Face Recognition Application for Security and Protection Systems -
Part 2: Face Image Data
ITU-T H.264 SERIES H: Audiovisual and Multimedia Systems, Infrastructure of
Audiovisual Services - Coding of Moving Video - Advanced Video Coding for Generic
Audiovisual Services
3 Terms and Definitions
What is defined in GA/T 893, and the following terms and definitions are applicable to
this document.
3.1 Video Surveillance
Video surveillance refers to a means of using video technology to surveille and record
the target.
3.11 Miss Alert of Watch List
When performing passage test on the test subjects on the watch list, if the system fails
to correctly alert during the whole process, then, it is a miss alert of watch list.
3.12 Miss Alert Rate of Watch List
Miss alert rate of watch list refers to the ratio of the total number of passage tests with
miss alert of watch list to the total number of passage tests of the test subjects already
on the watch list.
3.13 False Alert of Non-watch List
When performing passage test on test subjects that are not on the watch list, if the
system alerts, then, it is a false alert of non-watch list.
3.14 False Alert Rate of Non-watch List
False alert rate of non-watch list refers to the ratio of the total number of passage tests
with false alert of non-watch list to the total number of passage tests of the test subjects
not included in the watch list.
3.15 Response Time
Response time refers to the time interval from the moment that the test subject enters
the recognition region to the first alert issued by the system in the passage test where
an alert is generated.
4 Basic Structure of System
The video surveillance face recognition system in the security protection field
(hereinafter referred to as the system) is composed of two parts: registration and
recognition. The registration part is mainly composed of the face registration unit, as it
is shown in Figure 1 a); the recognition part is mainly composed of three parts: the face
image acquisition unit, the face comparison unit and the result analysis unit, as it is
shown in Figure 1 b).
main area of the face, the system shall detect the location of the face and obtain the
face image.
5.3 Face Comparison
The system shall be able to extract face feature from the acquired face image and
compare it with the face feature of all targets on the watch list to generate a similarity
value.
5.4 Result Analysis
The system shall be able to analyze the similarity value generated by face comparison
and output an alert message in accordance with the set threshold.
5.5 Alert Output
5.5.1 In accordance with the alert message, the system shall output the alert scene
image, the scene face image, the target’s face image on the corresponding watch list
and the related information.
5.5.2 The system should output the similarity value between the scene face image and
the target’s face image on the watch list.
5.6 Watch List Management
5.6.1 The system shall have the functions of querying, adding, modifying and deleting
the watch list.
5.6.2 The system shall have the functions of querying, adding, modifying and deleting
the target’s face image on the watch list and the related information.
5.7 Alert Record Management
5.7.1 The system shall be able to implement real-time recording of alert scene image
marked with the location of the target’s face, the target’s face image on the
corresponding watch list, and the related information, the similarity value and alert time.
The alert time shall include: year, month, day, hour, minute and second; the year shall
be in millennium notation.
5.7.2 The system shall have the functions of querying, conducting statistics on and
exporting alert records.
5.8 Parameter Settings
5.8.1 The system shall have the function of setting the comparison threshold
parameters.
5.8.2 The system shall have the function of setting the maximum length of the target
7 Test Methods of System
7.1 Design Principle of Test
7.1.1 Before the test, a detailed test scheme shall be formulated. The test scheme shall
be specified in terms of test environment design, test subject composition, test subject
passage mode and test procedures to guide the test.
7.1.2 The passage mode of the test subject shall simulate the typical passage mode
of the people using the system in the application scenario.
7.1.3 The test should adopt the mode of video recording, and the application field test
may be supplemented. The format of the video recording should comply with the
requirements of GB/T 25724-2010 or ITU-T H.264.
7.1.4 Under the circumstance of using video recording for tests, it shall be ensured that
the playback speed of the video recording is the same as the recording speed.
7.1.5 When the mode of video recording cannot be used for tests, a test platform may
be built in the same field environment, and the video source for multiple tests shall be
consistent.
7.2 Functional Test Methods
7.2.1 Face registration
The test steps are as follows:
a) Select a face image that complies with the requirements of 4.1 in GA/T 922.2-
2011 as the test image. Utilize the static face registration operation mode
provided by the system to perform the registration operation. After this
operation is completed, check the post-registration information directly from
the database or by utilizing other modes provided by the system;
b) The tester shall remove all the ornaments that may obscure the main area of
the face, stand in front of the camera used by the system and look at the
camera lens. Adjust the focal length of the camera, so that the camera can
focus on the tester’s face. In addition, ensure that the distance between the
tester’s eyes in the video image collected by the camera is not lower than 60
pixels. Then, utilize the operation mode of collecting and registering face
image data on the surveillance site provided by the system to collect a frame
of video image as the test image and register it. After this operation is
completed, check the post-registration information directly from the database
or by utilizing other modes provided by the system;
c) Select a folder that contains more than one pieces of face image data that
comply with the requirements of 4.1 in GA/T 922.2-2011 as the test folder.
Utilize the operation mode of importing and registering face image data in bulk
provided by the system to register all face images in the test folder at one time.
After this operation is completed, check the post-registration information
directly from the database or by utilizing other modes provided by the system.
7.2.2 Face image acquisition
The test of the face image acquisition function shall be carried out in accordance with
the following steps:
a) The tester shall remove all the ornaments that may obscure the main area of
the face, stand in front of the camera used by the system and look at the
camera lens. Adjust the focal length of the camera, so that the camera can
focus on the tester’s face. In addition, ensure that the distance between the
tester’s eyes in the video image collected by the camera is not lower than 60
pixels;
b) Initiate the face image acquisition function of the system; observe whether the
system correctly marks the location of the tester’s face in the real-time
displayed video image or observe a screenshot of the tester’s face image in
the system.
7.2.3 Face comparison
The test of the face comparison function shall be carried out in accordance with the
following steps:
a) Utilize the face registration function to register the tester’s face image in the
system database;
b) Set the comparison threshold to the value recommended by the system;
c) Conduct a passage test on testers who are already registered; check whether
an alert message is output.
7.2.4 Result analysis
Persons who are not registered in the system shall be subject to the following passage
test as testers:
a) After setting the comparison threshold to 0 in the system, perform a passage
test; check whether an alert message is output;
b) After setting the comparison threshold to the recommended value in the
system, perform a passage test; check whether no alert message is output.
7.2.5 Alert output
a) Utilize the operation mode of parameter setting provided by the system to set
the comparison threshold parameter to 0; set the maximum length of the
target list in the alert message to 1;
b) Utilize the face registration function to register more than one person’s face
images in the system database;
c) Perform a passage test on testers who are already registered; check whether
the alert message contains the information of only one target person;
d) Set the maximum length of the target list in the alert message to 2;
e) Perform a passage test on testers who are already registered; check whether
the alert message contains the information of two target persons;
f) Set the comparison threshold parameter to the value recommended by the
system;
g) Perform a passage test on testers who are not registered; check whether no
alert message is output.
7.2.9 User management
Utilize the operation mode of user management and authority management provided
by the system to establish multiple users and set different authorities for the tester.
After the tester logs in to the system as users with different authorities, check the
operating authorization.
7.2.10 System log
The test of the system log function shall be carried out in accordance with the following
steps:
a) The tester logs in to the system and performs more than one passage tests.
After completing the passage tests, the tester shall log out from the system;
b) Utilize the operation mode provided by the system to check the system’s
running logs and operation logs;
c) Utilize the operation mode of querying and exporting running logs and
operation logs provided by the system to respectively perform the operation
of log querying and exporting. After each operation is completed, check the
operation result.
7.3 Performance Test Methods
7.3.1 In accordance with the length of watch list supported by the system, the system
performance test may be classified into 5 types. The length of watch list corresponding
...
|