Unlock instant, AI-driven research and patent intelligence for your innovation.

Neural speech transliteration method based on distinctive features

A differentiated and neural technology, applied in speech analysis, speech recognition, natural language translation, etc., can solve the problems of scarce and rigid transliteration work, and achieve the effect of shortening the length and making it easier to understand

Pending Publication Date: 2022-05-13
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, transliteration of entire texts, especially speech-to-text, is extremely rare
At present, the method of transliteration is mainly based on rules, mapping English sequences to Chinese character sequences through the comparison of English phonemes and Chinese pinyin, or directly mapping English words to Chinese characters according to the written rules. However, this mapping method is Some of the generated text is more rigid, and they also deal with text

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

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Neural speech transliteration method based on distinctive features
  • Neural speech transliteration method based on distinctive features
  • Neural speech transliteration method based on distinctive features

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0023] like figure 1 As shown, a neural phonetic transliteration method based on discriminative features includes the following steps:

[0024] Step 1: Process the dataset to convert English text sequences of text labels to IPA sequences. The sentences before and after conversion are expressed as Figure 4 shown. Use this dataset to train the neural speech recognition model, and the neural speech recognition model can use the transformer model.

[0025] Feature extraction is performed on the waveform file of the audio signal of the speech segment, and the feature extraction method may adopt MFCC (Mel-frequency cepstral coefficients), or other audio feature extraction methods.

[0026] After feature extraction of the audio signal, the obtained feature value is a two-dimensional array, wherein the first dimension represents time and the second dimension rep...

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 relates to a neural speech transliteration method based on distinctive features, and belongs to the technical field of speech recognition. According to the method, firstly, an end-to-end voice recognition technology is used for recognizing audio, and then regularity conversion is carried out on a recognition sequence based on distinctive features. Wherein the end-to-end speech recognition part receives English speech and recognizes a corresponding international phonetic symbol IPA sequence; according to the rule part, an English IPA sequence is converted into a closest initial consonant / vowel IPA sequence of Chinese pinyin according to distinctive characteristics, and then the IPA sequence is converted into a pinyin sequence according to a mapping rule of IPA and pinyin. According to the method, distinctive features and an end-to-end speech recognition technology are combined, so that the method not only is helpful for transnational communication and English learning, but also can provide fun for people in life, and has good practical applicability.

Description

technical field [0001] The invention relates to a neural voice transliteration method based on distinctive features, and belongs to the technical field of voice recognition processing. technical background [0002] With the rapid development of computer technology, computer technology has been applied to all fields of society. At the same time, problems such as difficulty in processing massive voice data and difficulty in human-computer interaction have also arisen. [0003] The goal of speech recognition is to automatically convert human speech into text by computer. With the further development of artificial intelligence and deep learning technology, speech recognition technology has made significant progress. Most of the existing speech recognition technologies use methods based on deep learning. Among them, the distinguishing features are various features that can distinguish language units based on the natural features of speech. Usually a phoneme is labeled as "+" o...

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
IPC IPC(8): G10L15/26G10L15/02G10L15/08G10L25/24G10L25/30G06F40/40
CPCG10L15/26G10L15/02G10L15/08G10L25/24G10L25/30G06F40/40G10L2015/025G10L2015/086
Inventor 郭宇航王志鹏陈朔鹰
Owner BEIJING INSTITUTE OF TECHNOLOGYGY