AI intelligent voice stream acquisition

A technology of intelligent voice and voice streaming, applied in voice analysis, voice recognition, instruments, etc., can solve the problems of machine users' language understanding deviation, semantic understanding deviation, expression error, etc.

Pending Publication Date: 2021-06-04
浙江思考者科技有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] In the process of man-machine dialogue, the user will inevitably make mistakes in expression, which will lead to deviations in the machine's understanding of the user's language. At this time, the error correction mechanism is very important for the machine. Without this mechanism, the user will need to spend a long time Explain its intention clearly, and the corresponding user experience will be very bad. On the other hand, although the voice can be recognized very well, it cannot understand the purpose of your dialogue, and there is a deviation in semantic understanding.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0030] AI intelligent voice stream collection, including the following specific usage methods:

[0031] Step 1: Deployment configuration:

[0032] First download the latest python version in the win7 system environment, then install the download package python-3.6.4-amd64.exe (installation version) with JDK1.8 and Eclipse4.6 versions, and then add the installation directory to the environment variable for the device, Just add the python installation directory to the pth system variable;

[0033] Step 2: Deployment test:

[0034] To test whether the python installation is successful, open the cmd command line and enter the python command, enter print ('HelloWorld'), after the above tests, the python environment is installed;

[0035] Step 3: Connect to the cloud server:

[0036] The technical modules required by the AI ​​speech cloud server have four parts, namely automatic speech recognition, natural language understanding, natural language generation and text-to-speech;

...

Embodiment 2

[0046] Further, in the above solution, the core of the interaction process of the four module groups in step 3 is the intent.

[0047] Further, in the above solution, the intent is mainly expressed as the voice spoken by the user after using the AI ​​voice stream to collect, and these intents represent the core functions of AI intelligence. If AI intelligence successfully recognizes the user's intent, AI voice is required. After the user's voice is collected in the stream, the result is fed back to the user. If the AI ​​intelligence cannot identify the user's intention, it needs to give the user a prompt to guide the user to use it.

[0048] Further, in the above solution, how AI intelligence recognizes the de-intent is mainly to perform semantic analysis on the user's voice collected from the AI ​​voice stream, that is, to analyze and understand the voice recognition results. One machine instruction, then analyze the machine instruction, make a feedback response to the cloud ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses AI intelligent voice stream acquisition, a program is started, whether the acquired voice stream is sent at the moment or not is noticed, a sentence is spoken to a microphone, whether the voice stream is sent or not is checked, the acquired voice stream is converted into an array, and then whether the volume of the sentence is higher than a preset value or not is judged. And if the voice flow is higher than the preset value, the voice flow is transmitted to a cloud server for subsequent voice recognition machine logic processing. According to the invention, the voice flow of the microphone is monitored in real time, a user speaks at the microphone, the system collects the voice flow through the sampling frequency, converts the collected voice stream into an array, then judges whether the volume of the voice is higher than a preset value, and transmits the voice stream to the cloud server if the volume of the voice is higher than the preset value for subsequent voice recognition machine logic processing. According to the technology, the voice flow can be acquired in real time, and the voice flow is processed, so that the voice flow is sent in a more humanized manner.

Description

technical field [0001] The invention relates to the technical field of voice stream collection, in particular to AI intelligent voice stream collection. Background technique [0002] In the process of human-machine dialogue, the user will inevitably make mistakes in expression, which will lead to deviations in the machine's understanding of the user's language. At this time, the error correction mechanism is very important for the machine. If this mechanism is lacking, the user needs to spend a long time. Explain its intention clearly, and the corresponding user experience will be very bad. On the other hand, although the speech can be recognized well, it cannot understand the purpose of your dialogue, and the semantic understanding is biased. [0003] In the prior art, especially when the voice stream is collected and sent in real time, there is no trigger condition, which consumes a lot of network resources, but this technology can realize the real-time collection of the v...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & AuthorityApplications(China)
IPC IPC(8): G10L15/18G10L15/22G10L15/30G10L15/34
CPCG10L15/22G10L15/1822G10L15/34G10L15/30G10L2015/223G10L2015/225
Inventor谢伟平柯玉蒙
Owner浙江思考者科技有限公司