Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Speech recognition method and system based on end-to-end deep learning model

A technology of deep learning and speech recognition, applied to speech recognition with small vocabulary. , based on the end-to-end deep learning model in the field of speech recognition, which can solve problems such as large vocabulary and training time

Active Publication Date: 2019-01-04
BEIJING AIYISHENG TECH CO LTD
View PDF6 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But this method requires a large vocabulary and training time

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
  • Speech recognition method and system based on end-to-end deep learning model
  • Speech recognition method and system based on end-to-end deep learning model
  • Speech recognition method and system based on end-to-end deep learning model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0017] Embodiments of the speech recognition method based on the end-to-end deep learning model of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art would recognize that the described embodiments can be modified in various ways or combinations thereof without departing from the spirit and scope of the invention. Accordingly, the drawings and description are illustrative in nature and not intended to limit the scope of the claims. Also, in this specification, the drawings are not drawn to scale, and like reference numerals denote like parts.

[0018] The speech recognition method based on the end-to-end deep learning model of the present embodiment includes the following steps:

[0019] Step S10, classify and encode the finals and initials, map the finals with similar pronunciation to the same code, and map the initials with similar pronunciation to the same code to form a regular mapping table. The following T...

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 provides a speech recognition method and system based on an end-to-end deep learning model. The method comprises a step of mapping simple or compound vowels with similar pronunciations to the same code and mapping initial consonants with similar pronunciations to the same code to form a rule mapping table, a step of performing data encoding on a corpus by using the rule mapping ruleand representing Chinese characters of the corpus by using codes in the rule mapping table, a step of training the encoded corpus by using a hybrid end-to-end model to obtain 'pinyin' and 'phoneme' acoustic models, wherein the hybrid end-to-end model comprises a 'pinyin' end-to-end model and a 'phoneme' end-to-end model, a step of encoding multiple words to be applied by using the rule mapping table to form a vocabulary library, a step of identifying speech by using the acoustic models, and a step of comparing codes outputted by the acoustic models and codes of vocabularies in the vocabulary library by using editing distances and finding a minimum editing distance, and a corresponding vocabulary is an identification result. According to the method, the recognition efficiency of the systemis improved.

Description

technical field [0001] The present invention relates to the field of speech recognition, in particular to a speech recognition method and system based on an end-to-end deep learning model, especially suitable for speech recognition with a small vocabulary (such as limited commands). Background technique [0002] Voice, as an important way of human-computer interaction, has attracted more and more attention. Based on the current development status of speech technology, speech recognition scenarios with limited commands are the most mature and important, especially in some fields of interaction with information systems, where command-style interactions are clear and error-free. Especially in the medical field, a doctor's time is precious, and every minute saved may save an additional life. Speech recognition can greatly improve the efficiency of doctors using the information system and save the doctor's time in system interaction to the greatest extent. In addition, in some ...

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 & Authority Applications(China)
IPC IPC(8): G10L15/06G06N20/00
CPCG10L15/063
Inventor 赵明
Owner BEIJING AIYISHENG TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com