USPOREDBA OSOBA

CENT.SI

We are in the biometric enrolment business for more than 15 years. With our incredible simple, yet very powerful software solutions, we managed to gain trust from the governments, institutions and enterprises looking for highest and most reliable biometric enrolment solutions.

 Our biometric enrolment applications are used for enrollment of biometric data for national and international documents, such as biometric passports, electronic IDs, health insurance cards, visas and residence permits as the solution is built on top of ICAO, FBI and ISO standards. We are committed to bring the best end results. That is why we are working with world most recognized hardware providers. We can integrate your favorite camera for facial biometrics, your selected fingerprint reader for fingerprint enrolment and any signature table on the market. Our team of engineers provides consulting, development andintegration into existing systems

Pri svojem delu uporabljam programsko opremo podjetja Nec z nasledjimi specifikacijami:

Real Time Facial Recognition System (RTFS) Specifications:

1 - Collection & Matching

• The system must automatically locate a face from a CCTV video feed. The supported cameras should include:

• Avigilon HD2MP Pro

• Avigilon HD16MP

• Avigilon 16H4 PRO

• Axis – All cameras within the Axis range including WDR devices

• Bosch NDC-284-P

• Panasonic WV-NP502

• Sanyo VDC-HD3500P (2MP)

• Sony SNC-VB630

• G6-V Range

• The RTFRS system must support H264 RTSP and MJPG video streams

• The RTFRS system must support full color video for biometric matching and display purposes

• The RTFRS system must support a minimum of 60 pixels between the eyes

• The RTFRS system must support a minimum resolution of 1024*768 and must be able to make full use of resolutions up to 1080p

• The RTFRS system must support video capture from 8 to 30 frames per second

• The RTFRS system must support a horizontal rotation of -15 to +15 degrees, a vertical rotation angle of -10 to +10 degrees and a sideways rotation of -15 to +15 degrees

• The RTFRS system shall be capable of capturing a face from anyone 16 years of age and upwards, independent of ethnicity and gender

• The RTFRS system shall offer the ability to set thresholds based on a zonal level, or for a specific watch list or for a specific user

• The RTFRS system shall offer the ability to set camera configurations at a zonal level or for a specific camera

• The RTFRS must automatically crop the face

• The RTFRS system must provide the ability for a view of the CCTV feed which overlays a box around a subjects face and displays their name. The box should be color coded according to the subject’s watchlist. This should happen in near-real time (<5 second delay from the live feed).

• The RTFRS system must provide the ability to draw a box around faces in the scene which are not on the database (ie, unknown faces).

• The face must be automatically submitted to be 1:n matches against the RTFRS system

• The RTFRS system must provide a match score based on its internal algorithm

• The RTFRS must provide a response within under 5 seconds (network overhead excluded) against a database size of 100,000 faces.

• Any matches which score greater than a pre-defined threshold shall raise an alert

• Alerts shall either be processed automatically (in the case of a single hit over a pre-defined threshold) or sent for human adjudication (in the case of multiple this over a pre-defined threshold)

Registration

• The system must support the ability to bulk enroll records as an initial one time process. All enrollment images from shall be of ICAO quality.

• The system must include the ability to register subjects directly within the RTFRS, including:

• Photograph capture via webcam

• Photograph import via file

• Via a smart device (IOS / Android) app connected to the RTFRS via WiFi or cellular link

• All photographs can be validated to ensure they meet minimum quality thresholds, including checks such as:

• Face roll

• Pixels between the eyes

• Biographic data entry including:

• First Name

• Middle Name

• Last Name

• Date of Birth

• Gender

• Address

• Custom Field 1, Custom Field 2, Custom Field 3, Custom Field 4, Custom Field 5

Adjudication:

In the event that the system finds multiple matches over a pre-defined threshold, the candidate list shall be sent to an operator for adjudication. The adjudication system shall enable an operator to view the probe image (ie, the image captured from the CCTV camera) to view all potential candidates above a predefined threshold returned by the RTFRS system. The operator can then select a match/no-match against each of the candidates in order to validate the person’s identity. In order to aid the operator’s decision, the adjudication station should enable the operator to view the probe alongside any given candidate and carry out basic techniques such as the ability to zoom and rotate each image. Based upon the operators decision, if a match is found an appropriate alert will be raised. The adjudication station shall automatically be populated by the RTFRS system to enable an operator to view an active queue of pending adjudications. The queue shall automatically update based on an operators decision. Multiple operators should be able to connect simultaneously.

Alerts In the event that the RTFRS system raises an alert, the system shall have the ability to send an alert via:

• Email (via a NARA supplied SMTP server) to a predefined address

• On-Screen pop-up

• To a smart device (IOS / Android app)

Facial Accuracy

The RTFRS system must have been submitted for independent testing and must be able to independently demonstrate the following levels of performance.

1. Must have participated in NIST 2013 FRVT and NIST 2010 MBE tests

2. Must demonstrate a rank one miss rate of 3.1% or lower on a database of 160,000 records with good quality images (as used by NIST in FRVT 2013)

3. Must demonstrate a rank one miss rate of 7.9% or lower on a database of 160,000 records with low quality images (as used by NIST in FRVT 2013)

System Configuration

• The RTFRS system must be hosted on Windows 8.1 Professional or Windows 2012 Server R2, both in English language.

• The RTFRS system database must be Microsoft SQL Server 2012 or 2014.

• The RTFRS system must use standard Intel Xeon v3 CPU’s or Intel i7 4th Generation CPU’s and must fully support multi-core processing and hyper threading.

• The RTFRS system must use a Graphical Processor Unit (GPU) to provide facial extraction and must fully support either the Nvidea GTX 780Ti, Quadro K5200 or Tesla K20/K40.

• The RTFRS system must not require any proprietary hardware.

Dodatna strokovna znanja po potrebi pridobim tudi pri podjetju Visionmetric ltd.

3D-EFIT is a powerful and user-friendly addition to the E-FIT and EFIT-V applications which can rapidly deliver effective 3D face models from the 2D composite image. Witness stimulus and feedback is enhanced by the 3D view and allows them an unprecedented level of control over the image they are visualising. Features may be viewed from all angles and modified in real-time.

http://www.visionmetric.com/gallery/

Smo podjetje s trdnimi temelji in pomembnim položajem na našem trgu. Približno 10 let se že ukvarjamo z obnovo vaših podatkov in moramo priznati, da smo resnično dobri v tem.

Ponujamo vam:

• Obnovitev podatkov iz trdega diska (katera koli vrsta, operacijski sistem in vrsta škode)

• Obnovitev podatkov na medijih kot so: ZIP, LS, MO, JA

• Trajni izbris podatkov itd.