Facial Recognition: The State Of The Art

Face Recognition for CCTV
Face Recognition for CCTV The UAE has already approved facial recognition in a few sectors. This technology is rapidly enhancing, improving, and expanding globally. Increasingly adopted, facial recognition solutions help their users to make cashless payments, stay safe and secure in varied places, enter facilities, and check in. Let’s dive into the technology’s state of the art!

Facial Recognition In The UAE And Globally: Quick Facts

  1. The accuracy of the state of the art facial recognition systems is up to 99,9%
  2.  Facial recognition search results can be attained in a fraction of a second. Those numbers are claimed by reputable organizations like Japan’s NEC and the US National Institute of Standards and Technology (NIST).
  3. Biometric facial recognition is adopted in the ‘UAE Pass’
In March 2021 the Hamad Obaid Al Mansoori, Head of the UAE Digital Government and Director-General of the Telecommunications and Digital Government Regulatory Authority, said that –
“…the use of facial recognition technology in the registration process of the Digital ID app is a key step towards the implementation of emerging technologies based on artificial intelligence (AI), to establish a digital lifestyle in the UAE.”
4. Abu Dhabi police use a facial recognition system Abu Dhabi police enhanced their vehicles with a real-time biometric facial recognition system. This system is also implemented at the border checks. 5. Dubai public transport is secured with AI-powered facial recognition technology Facial recognition at Dubai Metro stations to identify wanted criminals 6. Facial recognition market size value is USD 4,45 billion in 2021, and the revenue forecast is USD 12.11 billion in 2028, while a CAGR growth of 15.4% is expected in 2028 (by https://www.grandviewresearch.com/industry-analysis/facial-recognition-market).

What Is Facial Recognition?

Facial recognition is software that:
  • captures
  • analyzes
  • and recognizes
– a person on a photo or video. That technology is one of the most powerful surveillance tools ever adopted by varied business domains.

How Does Facial Recognition Work?

The market offers plenty of facial recognition AI solutions, and every system works differently. Still, there’s a common 3-step process all of them follow:

Step 1. Detection (or Capturing)

This process involves detecting a face on an image or video. Most probably, you’ve already seen that in action, if you ever used a camera with a face capturing feature – it detects a face and draws a box around it then.

Step 2. Analysis (or Attribution)

This phase is meant for mapping faces detected in the previous step. The algorithm measures:
  • the distance between eyes,
  • the distance between nose and mouth,
  • the shape of the chin.
After that, it converts all this data to a so-called ‘faceprint’ – a string of points or numbers.

Step 3. Recognition (or Matching)

Finally, the facial recognition algorithm tries to recognize the face and confirm the identity of the person from the image or video, attempting to find the answer to the question “Who’s on that image/video?”

Sectors Already Disrupted By Facial Recognition

  • Transportation
  • Banking
  • Education
  • Gaming
  • Healthcare
  • Government services
  • Voting
  • Law enforcement
  • Travel & hospitality
  • Retail
  • Marketing & Advertising
  • Social Media

Facial Recognition Use Cases

Law enforcement (Security)

Facial recognition AI has already gained traction in varied security applications. Combating crime led to increased activity on this market. Facial recognition software brought the obvious value of detecting and preventing crime. Here are a few cases of how the technology is used in law enforcement:
  • Issuing identity documents – facial recognition is mostly used in combination with biometric technology (fingerprints) to prevent identity theft and other ID-related frauds.
  • Face match while border checks – facial recognition is used to compare the biometric passport photo with its holder’s face more often today. Police use the same methods of personal identification.
  • Drones with aerial cameras for facial recognition – mass events in large areas must be safe and secure, thus drones with facial recognition capabilities help to detect any disorders and react on time.
  • CCTV systems with facial recognition – help to keep public security:
  • support investigations,
  • find missing people,
  • identify and track offenders.

Financial sector

Have you heard about KYC (aka Know Your Customer) procedure? Facial recognition technology simplifies financial institution’s new customer onboarding online with its liveness detecting feature. Since digital account opening (DAO) got to the TOP5 trends of 2020 in banking, according to Forbes, face recognition and liveness detecting prove useful to the financial sector.

Retail

KFC, the American fast-food chain, Alibaba, the Chinese giant retailer, Amazon, the multinational technology company – what could they possibly have in common? All of them already adopted selfie payment! Selfie payment is the way to pay for products you buy, using your face. It employs facial recognition technology to recognize the payer, and automatically charge him or her the necessary amount for purchased goods.

VIP recognition

Staff can’t recognize valuable customers, VIP guests, and visitors of commercial entities immediately. Smart cameras with facial recognition app allow identifying them in a fraction of a second and provide VIPs with the right service at the right time.

Access control verification

Dubai Airport’s ‘Smart Tunnel’ is empowered with AI facial recognition, allowing controlling access to the airport, and verify visitors’ identities: Dubai Airport Smart Tunnel is the future of travelling

Facial Recognition AI: What’s on the market?

Pulsologic – beyond facial recognition! 

Pulsologic (by Innohub) is a contactless AI-powered system to measure human heart rate variability and reveal internal human states (like stress level) using computer vision technologies. The issues of other existing facial recognition solutions:
  • Existing technologies and known non-contact methods, like face detection and recognition applications, cannot reveal internal human emotions from neutral faces.
  • For security workers at border control (e.g. interrogation rooms) and police, it is crucial to swiftly identify criminal subjects trying to perform unlawful actions.
  • Regardless of all current training for the interrogators, it is extremely hard to pinpoint what the interrogated subject thinks of right at the moment. And given the interrogated person’s neutral face, it’s not clear to the interrogator, what this person feels at any given moment.
Pulsologic allows you to shed light on the feelings of the interrogated, even if he/she does not show absolutely any external emotions. Pulsologic can be used for:
  • Contactless remote HRV and stress tracking;
  • Person’s functional and emotional states detection;
  • Person identification & Authentication via remote detection of vital signs presence.
Pulsologic obtains a person’s:
  • Pulse;
  • RR-intervals;
  • Heart rate;
  • Breathing rate;
  • Stress level;
  • Productivity index;
  • Brain activity index.
Pulsologic can reveal a person’s:
  • Blood pressure;
  • Oxygen saturation;
  • Activation/Relaxation level;
  • Hidden inner feelings (levels of positive/negative emotions).

Possible Pulsologic applications

Case #1 Distant monitoring Effects are persistent to distances. For example, conference cameras work up to 35 meters. Case #2 Delinquency prevention The non-contact detection of criminals, for example, in a crowd of people. Case #3 Legal proceedings advancement The effectiveness of the passport and customs inspections, airports, and border control activities can be significantly enhanced using Pulsologic.

What’s Next?

If you are searching for AI or AIaaS solution that combines all the essential facial recognition features, including liveness detection, and emotion revealing, look no further than Innohub. Book your demo today! Facebook Twitter Youtube