In the past society, the ID card as the only certificate for people to enter and leave safely, because an ID card only corresponds to one person. But with the development of science and technology, the society is becoming more and more intelligent. The access certificate on the virtual network is not the ID card, but the only biological attribute that people have.
As long as it is the unique characteristics of people, it can be used to make a “key” on the virtual network, just like “only hearing its voice, you can know who you are” and “there are no two same leaves in the world” are all expressing this uniqueness. Nowadays, a variety of biometric technologies have been developed based on the uniqueness of fingerprint, face, iris, vein and voiceprint, such as fingerprint recognition, face recognition, iris recognition, vein recognition and voiceprint recognition.
In these biometric technologies, because of the popularity of smart phones, and now most of the smart phones have the function of fingerprint identification, and many companies take the way of fingerprint identification to count attendance, so fingerprint identification is the most widely recognized. But there is a kind of recognition technology which is not as popular as fingerprint recognition, but it has a unique charm in the field of science and technology, that is “voiceprint recognition”. Just as Professor Zhang cymbal, a leading scholar in the field of artificial intelligence and academician of the Chinese Academy of Sciences, said: there are limitations in current image recognition. With just a little interference, the machine will mistakenly recognize an “artillery position” as “a group of sheep”, or even anything else, which is very “fragile” and “unsafe”. At present, all kinds of biological features are relatively safe It’s all voiceprint.
But here we must make a distinction between “voiceprint recognition” and “speech recognition”. At present, the application of smart phones is speech recognition, and there is no really mature voiceprint recognition technology embedded in a mobile phone. The simplest distinction between the two is that the function of voiceprint recognition is to recognize “who is saying” while the function of speech recognition is to recognize “what has been said”.
See 100 billion application scenarios
At present, the application of voiceprint recognition is not very popular, mainly in the fields of public security, finance, social security and intelligent security.
National Public Security
Voiceprint is an important and difficult to imitate biological feature. Voiceprint identification technology is first used in the fields of national defense security, public security technical investigation, Judicial Correction and so on, which effectively guarantees the national and public security.
For example, in the battlefield environment, voiceprint recognition technology can detect whether there are key speakers in the process of telephone conversation, once the military command is issued through the telephone. The identity of the person who issued the order can be identified. It is reported that the US EP-3 reconnaissance plane, which was forced to land at Hainan airport on April 1, 2001, carried a voiceprint recognition and interception module similar to that of us. In the anti-terrorism war, the communication between terrorists before and after committing a crime often contains key contents. Therefore, the voiceprint recognition system is pre installed in the communication system or security monitoring system, which can prevent and track down criminals through communication tracking and voiceprint recognition technology.
Especially in recent years, the various forms of Telecom fraud. Criminal suspects emerge in an endless stream after calling the victim’s detailed personal information through illegal channels to blackmail, threaten and kidnap the victim through telephone. For example, in the “8.19 Xu Yuyu Telecom fraud case” which caused a sensation all over the country, the public security and judicial personnel can use the heating identification technology to lock the suspect from the voice and narrow the scope of criminal investigation. The voiceprint recognition system can be installed in public security checkpoints such as stations, airports and docks to effectively identify and prompt dangerous people, reduce the errors caused by naked eye recognition, and improve the safety of people’s lives and property.
Nowadays, online virtual currency payment has become the mainstream way of people’s transaction, so the identity authentication of online payment is becoming more and more important. Fingerprint recognition and face recognition have been exposed a variety of loopholes at the same time, voiceprint recognition has also been placed more expectations. In order to prevent theft and other cases, voiceprint recognition is added to the transaction payment, and dynamic password is used for client identity authentication to effectively improve the security of personal capital transaction.
From 2016 to 2018, China Construction Bank, Bank of Guiyang, Bank of Lanzhou, Bank of Xi’an, China UnionPay, China Mutual Financial Association and other units launched voiceprint identification services for account login, large amount transfer, card free withdrawal, password recovery and other business scenarios. WeChat and Alipay have also launched the login mode based on voiceprint dynamic password. In addition, the introduction of voiceprint recognition technology in the credit business as an anti fraud means can also effectively reduce the incidence of fraud and multi head loans.
Social security for many old people is the guarantee of their life. To get social security, the old people have to go to social security institutions to verify their identity in person, which is a hard and inconvenient thing for the old people. But it’s very important to verify your identity. How can you not need to go to the scene and guarantee that the social security fund will not be falsely claimed? Voiceprint recognition technology is an effective solution to this problem, but also for social security institutions to avoid a lot of human, material, administrative and time costs.
The pilot work of voiceprint survival certification has been carried out in many provinces and autonomous regions of China, aiming at providing convenience for people in many places.
Intelligent security and personalized voice interaction
Using voiceprint confirmation technology, it can complete the authorization of access control of various things in personal daily life, such as smart phone lock screen, voice password lock of various network accounts, computer voice lock, voice security door, car voice lock, etc; Using voiceprint recognition technology, it can support intelligent speaker, intelligent voice assistant, etc., provide personalized services for users of different ages and open specific functions and permissions, etc.; through the combination of voice recognition and voiceprint recognition technology, it can complete multi person meeting recording.
New application scenarios
According to the analysis of foresight Industry Research Institute, the current global biometric industry has a huge scale. The market scale of the segmentation direction of voice biometric (Note: voiceprint recognition) is nearly 10 billion US dollars, and it is expected to exceed 20 billion US dollars (134.6 billion yuan) in 2020, accounting for 2.4% of the whole biometric market.
From the perspective of network identity authentication application field, according to the data of foreign research institute markets and markets, the network security market is expected to grow to US $155.74 billion in 2019, of which the scale of identity authentication information security market will exceed US $30 billion. Voiceprint recognition market has become a booming blue ocean market.
Intelligent electronic devices must be a breakthrough in the application of voiceprint recognition. In the current electronic equipment, there is no representative and cross era voiceprint recognition products, such as fingerprint recognition or face recognition, the first cake in the application market is hopeless, and it is possible to be the first to eat the “voiceprint recognition” cake.
With 5g commercial further promote the construction of the Internet of things, Internet of things as the development goal, non-contact recognition technology voiceprint recognition will have a very big application scene. For example, in the field of smart home: voiceprint recognition will change the operating habits in the field of smart home in the future. There are many troubles in the operation and control of smart home applications, and voiceprint recognition is one of the effective solutions. Free the hands and feet of users, without walking around or holding the mobile terminal to press the button desperately. It also allows the elderly, children and other users who are not convenient to control the product system in physical form to enjoy the life of smart home, which is more conducive to the popularization and implementation of smart home in public families.
High technical barriers and serious polarization
The manufacturers of voiceprint recognition can be divided into three stages
The first stage: around 2000, the first batch of companies were established, mainly old voice manufacturers;
The second stage: around 2005, the second batch of companies were established, mainly specialized in voiceprint;
The third stage: the third batch of start-ups with AI labels, and in the past two years, bat and other large companies have begun to lay out their voice print field.
In November 2018, the people’s Bank of China issued the financial industry standard “technical specification for security application of mobile finance based on voiceprint recognition”, which means that voiceprint recognition technology has been recognized by financial regulatory authorities. The release of this standard also means that financial technology has become one of the most popular application fields of voiceprint recognition, and the scene application of voiceprint recognition in the financial field will greatly promote the development of upstream and downstream industries. At the same time, voiceprint recognition related manufacturers have been expanding rapidly, not only professional voiceprint manufacturers who have been working for many years, but also a large number of new faces have emerged.
As a result, the proportion of voiceprint manufacturers with clear technology sources is not high, and they are mainly concentrated in a few universities and scientific research institutions with deep cultivation.
Moreover, due to the complexity of sound signal, it is very difficult to process sound signal technology, which may lead to the old enterprises with accumulated technology to seize the opportunity of expanding application scenarios and develop better and better, while new enterprises may not be able to learn and master voiceprint recognition technology and grow up.
Since it’s so good, why isn’t it popular
Compared with other biological characteristics, voiceprint has the following three advantages
1. Non contact and easy to accept:
Because of its non-contact, voiceprint is a biometric technology that can be used in remote field. In addition, it is convenient and natural to obtain voice with voiceprint features. In the process of collection, less personal privacy information is involved, so it is easier to be accepted by users.
2. Low cost:
The voice acquisition device is low-cost and easy to use. A microphone is enough, and there is no need for additional recording equipment when using communication equipment (such as telephone and mobile phone).
3. Forgery is difficult
Voiceprint information must be collected in vivo, which can make voiceprint password change dynamically without worrying about password forgetting, losing and stealing, so as to prevent recording counterfeiting.
Everything that has to be mentioned has two sides. Advantages and disadvantages must exist at the same time. Voiceprint recognition also has some difficulties to overcome, such as the establishment of voiceprint database and features, far-field recognition, noise, multi-channel (mobile phone, telephone, network) and other problems need to be solved by the whole industry, none of these difficulties can make voiceprint recognition popular.
The future is worth looking forward to
Although voiceprint recognition technology has not been widely used, it is undeniable that it has been recognized by the times, given unlimited expectations, and the future can be expected. With the trend of interconnection of all things, it will also be the trend of interconnection of all kinds of data information, and ensuring the security of data information is related to everyone’s interests, so how to improve the security technology is an important topic in the future, and voiceprint recognition will be an important means to improve the security solution in the future. In the future, voiceprint recognition will play an important role and be widely used as fingerprint recognition and face recognition.
Editor in charge: CT