Fingerprint identification had once been the most widespread biometric technology around the world. However, many argue that this technology has quite a lot of shortcomings.
For instance, even expensive fingerprint reading scanners have a hard time identifying dirty or wet fingers, plus, some people's fingerprints are unreadable. Furthermore, being vulnerable to the temperature and precipitation, such scanners consequently cannot be used outdoors. Plus, fingerprint reading scanners do not meet today’s demand for contactless biometric technology.
According to a new comprehensive report 'Global Contactless Biometrics Technology Market 2020-2026', "the Global Contactless Biometrics Technology Market size is expected to reach $18.6 billion by 2026, rising at a market growth of 19.1% CAGR during the forecast period. The development and acceptance of contactless biometric technologies have been driven by demand for faster and easier authentication processes and boosted by demand generated by the COVID-19 pandemic." Thus, it is contactless biometric recognition technologies that meet the latest requirements.
Until quite recently, face recognition technology was too expensive and poorly scalable. Nevertheless, a lot of factors have changed in recent years. To start with, facial biometric technologies have become more accessible for a large audience. Being affordable, reliable, and easy to use, facial recognition systems provide a high level of security. Furthermore, the facial recognition system allows you to instantly notify about facial identification cases.The market of biometric technology is continuously growing
It is also important to emphasize that the system itself automatically updates biometric data. Photos in biometric profiles can be updated directly from the video stream. The data is stored in long-term storage and does not take up much memory. The reasons mentioned above provide all business fields with a competitive advantage. Since the market of biometric technology is continuously growing, contactless identification will be highly demanded in the long run.
Impact of COVID-19
Plus, the contactless facial recognition system is especially relevant today due to the COVID-19 pandemic. Now wearing a mask is required almost in all public places. That is why those systems aimed at people's safety monitoring had to promptly develop their solutions according to the new requirements. Developers of facial biometric solutions have encountered an issue of face detection in masks. It was essential to adapt the software to such changes, more specifically update the face recognition algorithm.
It may be illustrated by the case of RecFaces company. RecFaces developers have updated the facial biometric algorithm to ensure the most accurate recognition of people in masks that cover almost 50% of a person's face. Nonetheless, if the company forbids entering its territory without a mask, the system sends notifications (push or SMS notifications) to control people coming through the checkpoint with and without masks. The algorithm update has boosted face recognition accuracy and speed.
As a matter of fact, facial recognition algorithm has evolved around the world. According to the tests conducted by the National Institute of Standards and Technology (NIST), the top face identification algorithm of 2020 has an error rate of 0.08% compared to 4.1% for the best algorithm in 2014. Such improvements will reduce risks linked to misidentification, and expand the advantages that can come from proper use in the long run.Al and deep learning are key elements of the latest-generation algorithms
According to the National Institute of Standards and Technology report recognition errors were caused mainly by image quality variations like pose, illumination and expression. In 2018 the software was at least 20 times more accurate than it was in 2014 and in 2019 finding “close to perfect” performance by high-performing algorithms. Such improvement has resulted from the integration or replacement of previous approaches with those based on deep convolutional neural networks, operating even with poor quality images. Artificial Intelligence (Al) and, more specifically, deep learning are key elements of the latest-generation algorithms. Facial recognition is reaching that of automated fingerprint comparison, which has been considered as the gold standard for identification for a long time.
Therefore, there is no doubt that innovation drives the development of solutions, and biometric technologies also move with the times. The shift from fingerprints to facial recognition is a vivid example of such evolution!