How does facial Recognition System work?
Since the outbreak of COVID pandemic, we can observe the demand for adopting facial recognition system had been greatly increasing. The volume shot rapidly and believe to overtaking biometric fingerprint technology. We all knew that facial recognition system technology exists in the past 5 years, however, it’s not as popular compared to fingerprint systems in the overall marketplace before COVID-19. Today, we are going to discuss more how facial recognition becomes mainstream in identity verification for attendance, access control solutions.
Face recognition system adopt high-resolution imaging technology with resolutions of 2 megapixels. It’s used to detect live face whenever there’s person show up their face. In milliseconds, system capture persons face image and proceed to analyze with deep learning face algorithm. The system first converts the image into a digital string and establishes face print. Thereafter, the face algorithm will analyze based on transformed code from human face geometry, eyes, tone of skin, the distance between forehead and chin as well as face shape, ears, lips.
The next step is using a complex face algorithm to perform matching against the system database to find the right person. Lets’ talk a little about how the system handles the matching process. Nowadays, face recognition verification mode classify into two modes which are 1:1 and 1: N. Indeed, the 1:1 (one to one) mode is two factors authentication such as RFID Card with Face and 1: N (one to many) is single authentication which is using face only to perform matching against the database.
If the configuration is defining to 1:1, the face recognition system will first read the RFID card on the RFID card reader (1st authentication) and follow by capture the face to perform dual validation. Of course, the sequence can be vice versa depending on product manufacturer design workflow. Whereas for 1: N, the system will direct perform face comparison in the database. The system will display a person’s ID numbers, name, and photos if it’s successfully matched. Otherwise, it will prompt a rejection message or face not found.
1) What Factors affect face-scanning quality?
Undoubtedly, all this happens in milliseconds which is considered extremely fast. Most face recognition devices support scanning speeds between <0.2 to 3 seconds per user. However, some factors may affect speed such as environmental factors like exposure lux (sunlight and lighting). Of course, there are infra-red or supplementary lights to handle poorly lit environments to optimize scanning speed. As a result, some devices support better imaging technology like Wide Dynamic Range (WDR) which could help to reduce strong backlight.
2) How does it work with attendance software?
The authenticated will end by records timestamps and event logs into devices’ flash memory to serve attendance purposes. Normally the data can be exported using a USB thumb drive, TCPIP LAN, WiFi, or 4G/5G network. In addition, most of the face recognition reader provides client software and SDK to allow a software developer to get the access logs and tabulate into time attendance software. Thus, the attendance software will have based on received timestamps to determine staff clocking date and time. Then, the next steps are to calculate working hours, overtime, lateness, and others based on assigned working shift patterns.
3) How does the face recognition reader work with the access control system?
Face recognition system support third-party IO output control module. It’s a relay output to trigger signal to engage or disengage electromagnetic lock (EM Lock), turnstile system, lift controller, or vehicle arm barrier gate.
Door Access System
When adopting a face recognition system with door access, it can work with electromagnetic Lock, drop bolt, or letch lock. Moreover, it includes overwriting key switches to allow the user to deactivate the door system if the system is working abnormal or malfunction. On another side of the door will install with emergency break glass. Break glass is mandatory under the fire safety act to allow people in the building able to escape during an emergency evacuation. Besides that, the system also supports integration with the fire alarm panel. Other than break glass, an exit push button or touch-free sensor is use to allow people to trigger the door lock to disengage upon exit. In some circumstances, the user will install out the reader to enable capture of exit timing. It can be a slave card reader, PIN code, or face recognition terminal.
Turnstile and Barrier System
Similarly, face recognition support integration to tripod turnstile or any other kind of pedestrian, vehicle barrier gate. Well, We take tripod turnstile as an example. Tripod turnstile is a waist-high access control turnstile with three revolving arms. It uses to restrict and regulate entry/exit access in building at all times. Implementing this solution could boost operational productivity, and save the cost of manpower operation.
4) What’re the Benefits of using face recognition?
- Eliminate Time Thief compare to swipe card or PIN cause of buddy punching.
- Tighter Security through authenticate an employee’s facial features and grant him/her access permission.
- Time-Saving and Ease to Use, simplify enrolment process by just upload face photo into system.
- Seamless Integration of any third-party hardware and software application especially HRMS and Payroll Software.
- Reduced Contagion like use a finger scanner can easily transmit harmful viruses and bacteria between employees
- No worry to forgot to bring a card, your face is key to access
- Eliminate scanning issue happen on the dry, cut or wet finger.
5)What else other than using face recognition?
There are a few other types of biometric access control methods:
- Fingerprint Recognition System
- Palm Print Recognition System
- Iris Recognition System
- Handpunch System
- Vein Recognition System
6) What’re the important criteria for your consideration in buying a face recognition terminal?
- Understand product features. One of the major concerns is face accuracy, speed, and face spoofing.
- Determine who are the user such as staff, visitors.
- What function and features you need to meet your requirement. Examples like door access, two-way video, audio intercom, live view, and etc.
- Server attendance or access control?
- Whether the system deploys in a single location or multiple sites.
- Environment factors, like indoor, outdoor or partial outdoor. This will affect hardware durability and reliability.
- Cost budget
AP-T15
- Compatible indoor/outdoor door access EM lock system.
- Capture person attendance
- Max.1,500 faces and cards capacity
- Face verification speed ＜ 0.2 s/User.
- Accuracy rate ≥ 99%; IP65.
- 4.3-inch LCD touch screen, 2 Mega pixel wide-angle lens.
- Two-way audio with client software, indoor and master station.
AP-T60W
- Compatible indoor door access EM lock system.
- Capture person attendance
- Max.6,000 faces and cards capacity
- Face verification speed ＜ 0.2 s/User.
- Accuracy rate ≥ 99%; IP43.
- 4.3-inch LCD touch screen, 2 Mega pixel wide-angle lens.
- Two-way audio with client software, indoor and master station.
AP-T671
- Compatible indoor/outdoor door access EM lock system.
- Capture person attendance
- Max. 6,000 faces and cards capacity
- Duration ＜ 0.2 s/User.
- Accuracy rate ≥ 99%; IP65.
- Max.10,000 cards, and Max. 150,000 event capacity.
- 7-inch LCD touch screen, 2 Megapixel wide-angle lens.
- Two-way audio with client software, mobile app, indoor and master station.
- Alarm input and output capability.