Cutting: voiceprint recognition in westward journey !, Westward journey

Cutting: voiceprint recognition in westward journey !, Westward journey

Recently, the westward journey of the classic works of xingye has been released, leading to a nostalgic craze. Xiaobian also bought tickets and wanted to review the legendary one.

Unexpectedly, after watching it for a while, the editor did not control the wide-opening brain holes and fell into the question: in the movie, Zixia said, "open the door with sesame seeds ", the door of the wire hole is automatically opened, and the door is not opened, but it is also killed when Zun Bao says "Sesame opens the door.

I didn't understand why when I saw it ten years ago, but now I suddenly realized:

The door of the wire hole must use the voiceprint recognition technology!

Now the question is: what is voiceprint recognition technology? Is there any application in the real world?

Let's talk about it one by one.

Voiceprint recognition technology, also known as speaker recognition technology, is a biometric recognition technology that extracts voiceprint information from the speaker's voice signal and verifies the speaker's identity. Each person has a unique voiceprint, which is gradually formed by the growth of our vocal organs. No matter how similar others imitate our speech, the voiceprint is actually quite different.

Therefore, if your electronic device uses voiceprint recognition technology, it can only be opened by you. Now, all my friends know why Zixia's commands are the only way to get rid of his throat ~~

Of course, such cool technology does not only exist in film and TV, but has already entered our daily life.

In 2013, Tencent was the first in the industry to launch the screen lock application with voice and voice. After successful voiceprint registration three times on the device, you can use the voiceprint lock application. When the phone is locked, the user only needs to say a voiceprint password to the phone to unlock the phone. If not, the phone cannot be unlocked.

So how does a machine recognize our voice?

Voiceprint recognition has two key issues: feature parameter extraction and pattern matching.

The task of extracting feature parameters is to extract and select acoustic or language features that have the characteristics of strong separation and robustness to the speaker's voiceprint. In general, it is to extract the characteristic parameters that can represent the speaker's specific organ structure or habitual behavior from the speaker's speech.

This feature parameter is relatively stable for the same speaker and does not change with time or environment. At the same time, this parameter is consistent when the same speaker speaks different words, however, when different speakers say the same word, it can reflect significant differences and is difficult to imitate.

At the same time, the application of voiceprint recognition technology requires rapid response on any occasion and system. Therefore, voiceprint recognition technology also requires fast operation mechanism and high noise resistance.

Figure: Principles of voiceprint Recognition

Unexpectedly, in the case of electro-optic Firestone, the equipment has completed such a complicated identification process in the Process of accurately determining whether we are actually "Masters.

Of course, the current voiceprint recognition technology is not perfect. For example, when your physical status changes and someone else records your voice, the voiceprint recognition system may also be mistaken. In response to the above problems, Tencent is also actively researching solutions to help you quickly use secure, fast, and reliable voiceprint recognition applications and products.

I believe that with the efforts of xunfei, voiceprint recognition technology will certainly have a broader application space in the future. In the near future, you may log on to your bank account with your voice, sign a contract with your voice, and lock the door with your voice ...... Eliminate the complicated procedures and return to the most natural voice interaction. Our future life will be more convenient and colorful because of the wide application of speech technology!

More open capabilities, please stamp: http://open.voicecloud.cn/