Increasing demand for facial recognition systems

Commercial demand often fuels innovation. In the case of facial recognition systems, requests for security surveillance and marketing solutions are spiking. In fact, just in the US, the facial recognition industry is expected to grow from $3.2 billion in 2019 to $7.0 billion by 2024, as forecast in “Facial Recognition Market” by Component.

Example Services: Now and Near Future

• Security Monitoring Systems
• Ensuring Public Safety
• Online Digital ID Authentication
• ATM Authentication
• Personalized Digital Marketing
• Airline Boarding

How it works

Face recognition technology works on similar principles of the human brain. We first see a person, pay attention to the features of their face. We instantly process them in our brain to determine a recognition. A similar process is duplicated by technology. The system acknowledges a face, targets a specific area, and immediately starts analyzing the image.

Different Models

There are a variety of types of systems using their own proprietary algorithms. Sometimes the system determines the similarity of proportions, selects the contours in the image, and compares them with the outlines of the faces. Another technique selects symmetries using neural networks. Experts believe the Viola-Jones method is the most effective because it can be used in real-time. Such a system can recognize a person’s face even when turned 30 degrees. This method works on the signs of Haar, where black-and-white rectangular masks of various shapes get superimposed on the image. Then the algorithm adds up the brightness of the pixels of the image that are under the pixels of the mask and calculates the difference in values. The system compares the results with the downloaded data and continues to track the face in the image (while choosing the right angle and adjusting the image quality).

3D Scans

The program looks for known reference points on the face. It highlights about 100 of these points. The most important measurements include, the distance between the eyes, the width of the nostrils, and the length of the nose. 3D images make it possible to recognize an object better. S

Facial scars, pockmarks, and bone structures are used. This technique includes using several cameras photographing the person from different angles. This way, a three-dimensional model is created. 

Liveness Test

Traditional face recognition software is not always sufficiently reliable for authentication models. In order to minimize fraud attempts, participants are asked to prove they are real and alive. To do this, they get asked to smile, blink, or say something as cameras record. Such a system would be much more difficult to deceive while using a mask or printed face photograph.

Existing Applications

All over the world, authorities are making efforts to improve the safety and comfort of residents. For example, in the subway of Shenzhen, a travel payment system using such technology was launched. For its implementation, a 5G network was launched. Now people do not show a ticket. They simply go to the screen at the entrance of the station and have their face scanned. The system recognizes the person and automatically deducts money from the linked account. After scanning, people can read information about themselves (gender, age, length of stay in the subway) and see photos.

A similar practice is being introduced in the Moscow metro:  In 2018, they began testing the biometric identification system. With the help of cameras, police officers track wanted criminals.

The Japanese company NEC has launched the world’s first ATMs that issue money through face recognition. Face recognition technologies are also widely used at domestic and international airports. For example, Amsterdam Schiphol Airport uses face recognition devices during the boarding process. Together with Cathay Pacific, the technology is still being applied in test mode. After checking in for a flight, a person scans a passport, a boarding pass, and after that passes passport control. Before boarding, the passenger’s face is scanned. If the image matches the photo from the database, the system provides access. This service speeds up the boarding procedure.

The popular iPhone X also authenticates owners using a facial recognition algorithm. Using an infrared camera, light sensors and a point scanner help create a three-dimensional map of the face. In theory, the system recognizes a face even in low light. Users can be authenticated even with glasses and a hat, to help unlock the iPhone conveniently.

Global Facial Recognition 

Biometric facial recognition is still maturing. But business analyst predicts a very promising future for business applications. As face scanning becomes everywhere, cloud-based data sources will emergy. Gradually, there an enormous global database of 3D facial images. It will be used to search for missing persons, find criminals, and authenticate consumers. Facial recognition systems will begin to be embedded by default in mobile devices, ATMs, POS-terminals, cars, and many IoT devices.

Interpreting Emotions

The next step in facial biometrics will be the recognition of human behavior and emotions. Research is being done in this area, and some systems are currently live. AI algorithms have already learned to determine basic emotions. The goal is better to interpret facial expressions for a wider range of people. Over time, all challenges will be solved. Future applications would be enormous, including healthcare (patient pain levels), retail(advertising impact), and public safety(identifying risk).