Method, system, device and medium for automatically adding punctuation marks to text

A punctuation mark and automatic addition technology, applied in the field of speech recognition, can solve problems such as difficulty in meeting the needs of different fields, and achieve the effect of taking into account the effect and processing speed
CN112906348BActive Publication Date: 2022-04-26GUANGZHOU YUNCONG INFORMATION TECH CO LTD
6 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUANGZHOU YUNCONG INFORMATION TECH CO LTD
Publication Date
2022-04-26

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

A method, system, device and medium for automatically adding punctuation to text, obtaining text by recognizing audio; converting corresponding text into multiple index value sequences, and inputting the multiple index value sequences into a deep neural network model, Obtain the probability distribution of each index value sequence; determine the maximum probability distribution value corresponding to each word in the index value sequence based on the probability distribution of each index value sequence, as the index of the punctuation mark to be added after the word; pass the index Obtain the corresponding punctuation marks from the predetermined punctuation mark index table, and automatically add them to the text sequence to complete the addition of punctuation marks to the text; if the index corresponds to a blank label, the current word is skipped and no punctuation is added to the current word symbol. The invention can realize functions such as automatic punctuation, cross-domain transfer learning and radical adjustment, and can also change the radical of the deep neural network model to meet the requirements of accuracy and recall in different scenarios.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of voice recognition, in particular to a method, system equipment and media for automatically adding punctuation marks to text. Background technique

[0002] Speech recognition can transcribe speech into corresponding text, but since punctuation marks themselves do not have pronunciation, the transcription result of speech recognition is often a text without punctuation marks. Adding punctuation to transcription results with additional tools can increase the readability of transcription results. Especially in the transcription scenario of long audio, punctuation marks are more critical for humans to understand the content of long texts. Common automatic punctuation tools are implemented by training a deep neural network model.

[0003] However, the existing automatic punctuation tools often have the following defects:

[0004] 1) The independent punctuation model is often not aimed at the single scene of ...

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