With only one month left before the end of 2020, the domestic social order is quickly recovering and is on track. However, various fields still have to bear the irreversible impact of the epidemic and are facing difficulties in reshaping.
In terms of biometrics, fingerprint recognition was the most widely used before the epidemic, but during the epidemic, it was regarded as an "unsafe" biometric method due to its contact action; face recognition was the most convenient before the epidemic, but now it needs to be done first. Take off the mask, which is slightly cumbersome.
On the other hand, non-contact biometric methods such as vein recognition have prominent advantages in the epidemic, and it seems that they will gradually occupy a larger market.
Under the epidemic, will biometric technology face a reshuffle?
Fingerprint and face recognition still dominate
In the current biometrics, it mainly includes fingerprint recognition, face recognition, iris recognition, palmprint recognition, voice recognition, vein recognition and other recognition methods. Based on different principles and the maturity of the industry chain, fingerprint recognition and face recognition have been Occupy the mainstream position.
However, according to the latest report of ABI Research, due to the impact of the epidemic, smartphone shipments have declined, and contact identification carries the risk of spreading the new crown virus. It is estimated that by the end of this year, global fingerprint identification revenue will fall by 22% to US$6.6 billion. .
Since the development of fingerprint recognition, it has gone through many iterations. In 1998, Siemens launched the world’s first prototype with fingerprint recognition technology. Its sliding fingerprint recognition has been developed until 2013. It was not until the release of the iPhone5s mobile phone that the first push fingerprint recognition entered the market. Know. Nowadays, fingerprint recognition has already broken the limitations of the push type, and has developed in capacitive and optical types.
At the 2020 Biometrics Technology and Application Forum, Chongqing Huifan Technology Co., Ltd. CEO Gao Xin Wang said that fingerprint recognition has expanded from capacitive to optical and ultrasonic, and exists in their respective market segments according to their own characteristics and design requirements.
Gao Xinwang believes that compared to capacitance and optics, the biggest feature of ultrasound is that it is not affected by materials, can penetrate glass, metal, plastic and other materials, and is not restricted by dirty environments and wet fingers. Ultrasonic fingerprint recognition has a bright future. .
After smartphones are equipped with fingerprint recognition, face recognition immediately enters the consumer electronics field. It uses a camera to collect facial images or videos, and automatically detects or tracks the face in the image to complete the recognition. Face recognition has undergone the transformation from recognition type and confirmation type to watch list type. With the improvement of algorithms and systems, it has been widely used. The accuracy of face recognition technology with masks that has grown with the epidemic is continuing to rise.
The security of face recognition has always attracted much attention. According to the "Public Research Report on Face Recognition Applications (2020)", 60% of the respondents believe that face recognition is abusive, and 30% of the respondents said that they have been Privacy or property security suffered by information leakage or abuse. Facing a security risk, face recognition is changing from 2D to 3D.
Compared with 2D, which can only obtain two-dimensional images, 3D recognition can collect up to hundreds of thousands of information points to obtain depth information and prevent fake identity. At present, quite a number of mobile phones are equipped with ToF lenses to realize 3D perception. It is worth noting that the current face recognition is mainly used for higher accuracy and lower cost iToF, dToF is mainly suitable for longer distance measurement, such as rear cameras. 2.5 meter measurement.
At this Biometric Technology and Application Forum, Austin, the application engineer manager of Shenzhen Baiou Biometric Technology Co., Ltd., said in a dialogue with the media that the current mobile phones are mainly single-point ToF, which has a lower cost. With the array ToF The cost is lower, and its effects and applications will be more expanded.
Although fingerprint recognition and face recognition are inevitably affected by the epidemic, these two technologies are still developing. According to a recent data from Transparency Market, fingerprint recognition accounts for 58% of all biometrics. Face recognition Accounting for 18%, it is still the mainstream of biometrics.
Strong development of non-contact biometrics
The other side of the epidemic is stimulating the economic growth of non-contact biometrics, including vein recognition.
The latest report by Future Market Insights predicts that from 2020 to 2030, contact biometric technology will follow the wave of the new crown pandemic, with a compound annual growth rate of 17.4%, and a five-fold increase in the global market size to reach 70 billion US dollars.
Vein recognition is a way to identify specific wavelengths of light in flowing blood and analyze vein images in the background to identify identity. Vein recognition can be divided into finger veins, palm veins, and ocular veins, with ocular veins having the lowest market share. Vein recognition itself is not a new technology, but its application is far behind face recognition and fingerprint recognition.
At the end of September this year, Amazon launched the Amazon one payment device based on non-contact vein recognition, which attracted great attention. Vein identification is difficult to tamper with and embezzle, and it is the most widely used in the financial field, but generally requires large equipment to provide support.
A representative of Nie Zedong’s team from the Institute of Medical Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, said at the 2020 Biometric Technology and Application Forum that finger vein recognition is similar to fingerprint recognition. Both require the person to be identified to reach into the image capture device to obtain the image. Active authentication, the difference is that it can be identified without contact, with the help of the existing widely established fingerprint identification user habits, the promotion of vein recognition has great advantages.
"However, the current vein recognition products are still facing imaging and accuracy problems. 3D rotation and finger offset will cause recognition problems. As humans age, they will also affect them." Nie Zedong team representative said.
However, this does not hinder the growth of the vein recognition market. Related reports predict that by 2020, the market share of vein recognition will rise from 2% in 2015 to 8% of the market, with a market size of approximately US$2 billion.
In addition to vein recognition, according to GIA's forecast, the global iris recognition compound annual growth rate will reach 18.8% from 2020 to 2027, and the market size will reach 10.3 billion US dollars by 2027, and the compound annual growth rate of voice biometrics will be 17 %.
Multimodal fusion is the future trend
"In terms of ease of use, variability, and security, different biological characteristics have their own advantages and disadvantages. Compare fingerprints, veins, faces, and iris. The accuracy of fingerprint and face recognition is far lower than that of veins and iris. We predict that there will be more and more multi-modal biological integration in the future. For example, the processing of banking business will transition from the earliest authorized smart card to fingerprint plus smart card, and then continue to evolve into fingerprint plus facial recognition, or fingerprint plus iris. "Liu Haolei said.
Like the trend of sensor fusion, biometrics based on various sensor technologies are also moving towards multi-modal fusion.
Perhaps this is the real "reshaping" of biometrics at the moment and even after the epidemic.