Name: SMART Tracker FRS

Text: (c 4

Speech

E;;:pLOgy

SMART Tracker FRS
Face recognition solution for real-time biometric search
and facial D 6 management in law-enforcement and s,ecurity agencies
V~deosumelllance becomes more and more s~gn~f~cant
component In complex securlty systems as due to crtmlnal sltuatlon, terrorist threats,
vulnerabll~tyof strategLc facll~t~es
d~str~buted
v~deosurvelllance systems w ~ t hefflc~entv~deoanalyt~cscapabll~t~es
can be qu~teefflc~entIn
lnvestigatlve actlvltles.
By concentrating people at h ~ g hdens~tyIn access~bleplaces, at regular or pred~ctabletimes, mass gathertngs present the opportunity for
terrorists to lnfl~ctmass casualt~es,cause economlc damage, and lnstlll public fear.
Next-generat~onIntellectual v~deosurvelllance systems should be deployed m places of mass gathering Such systems may contaln "state-of~dent~f~cation
of the potent~allydangerous persons Most commonly "h~dden"
the-art" face recognltlon technolog~esallowlng real-t~meb~ometr~c
checkpoints can be found In the entry and exit zones of tram stations, stadiums, subways, airports and In any other place, where the speed of a
stream of people IS llmlted.
SMART Tracker FRS deployed In a real-t~memode has the following performance parameters: a dangerous person will be detected In 97 cases
of 100 In 1-3 seconds STC solut~onSMART Tracker FRS can be used as an Ideal tool for creatlon forens~cphoto-arch~vesB~ometric~dent~f~cat~on
procedure takes only 3 seconds w~thlnthe database contalnrng 1 mlll~oncards. Any source can be used to upload new fac~alpictures lnto
database.

APPLICATIONS:

ADVANTAGES:
Real trme face detect~onand tracking In video stream
Algor~thmsIndependence of aging, beard and hairstyle changes, skln color changes
Adaptat~onto llghtlng cond~t~ons
Image v~sual~zat~on
on the monrtor llve v~deoand ~ d e n t ~ f ~f aecd~ s
Nav~gat~on
In v~deoarchlve and database and saving of all the required parameters

Transportat~ontram stat~ons,marine and river
termlnals; subway, alrports
Recreat~onalfacll~t~es
Strateg~cfacll~t~es
City infrastructure

l

FUNCTIONS:
Face detect~on~nl~vestream
B~ometr~c
search In real tlme and on request
Notif~cat~ons
on operators' desktops, by e-mall.

-.SMS
. , .MMS
.. .Video streaming archives, linking of
~dent~f~cat~on
procedures wlth v~deo-record~ng
Database (up to several rnlll~onscards)
contalnlng fac~alplctures w ~ t hmetadata
Facial Images can be uploaded lnto a database
from any source. photographs, v~deo-flies,
photo-cameras, scanners, web-cameras, CCTV
cameras
Automat~cquaky assessment of uploaded
plctures by IS0 19794-5 standard
Full workflow automation, access r~ghts
Cmltatlons, user activity aud~t

TASKS:

-

Real-tlm d&pchon oT w m t d persans
Structured database of facd data and metadata
Search lri database by photographs and metadata

U

Search operator
IP camera

L

/-7
DB operator

lnvestiqative
Tracking station

Central server and
computer nodes

CHARACTERISTICS:
Customer convenience:
Fully automatic lmage preprocesslng
User-friendly Interface
Bullt-ln context help
Accuracy:
97% of correct matches in case of 1 % of false acceptance rate

100%correctly matched faclal ssmples In 1 m~lliondatabase In
case of analysls by the operator 10 best results on the database
according to IS0 19794-5 s t a ~ d a i d s
Performance:
ldentlflcation result in 1-3 seconds
Blometrlc search In 1 mllllon database In 3 seconds
Minimal lmage requirements for faclal recognition
Inter-pup11spaclng larger than 32 pxels
Pose up to +/-l 5 " devtatlon from frontal lmage

-

Batch input and Integration with the exlstlng databases
Failure tolerance:
Fully distributed architecture
Back-up at all levels and redundancy
Quick recovery and robustness to hardware rnalfunct~on

REQUIREMENTS
TO CONTROL POINT

Biometric database management:
Bullt-ln Image quallty assessment by IS0 19794-5 standard
Face detect~onin vldeo-recordings and group Images, subsequent
card creation automatically
Controlled Input of faclal Images thanks to the integration wlth
photo-cameras, web-cameras and scanners
Security
Group- and role-based access control to system functtonallty
User actlons logging on central server and user actlvlty audit tools
L ~ n kbetween system components based on H n P S protocol
Scalability:
Up to 12 cameras connected to 1 tracktng station
Up to 8 tracking stations connected to 1 computing node
Up to 30 computing node connected to 1 central server
Up to 500 users for 1 central server
Easy deployment:
Web-based interface IS enough for execution most of the activities,
no need to Install the system on PC
M~nlmurnrequlrements to network capaclty allow usage of the
exlstlng IT-tnfrastructure

Distance between camera and control zone is from 1.5 to 30
meters (depends on Lens used)
Control zone breadth (picture width): 2 meters
Max. quantity of faces in video stream: up to 5
Lightning conditions in control zone: 300-1000 lux

g3)

Speech

0g
l;y;:

Russia

Germany

4 Krasutskogo street
St Petersburg. 196084
Tei +7812 325 8848
Fax -7 812327 9297
Email stc spbBspeechpro corn

Saargemunderstr 21 1
0-661l 9Saarbrucken
Tei +49(0)681 8590565
Ernall ~nfo@speech-techde

Document Path: ["1313-speech-technology-center-brochure.pdf"]

e-Highlighter

Click to send permalink to address bar, or right-click to copy permalink.

Un-highlight all Un-highlight selectionu Highlight selectionh