Skip to main navigation Skip to search Skip to main content

Large Vocabulary Continuous Speech Recognition with Deep Recurrent Network

  • Harbin Institute of Technology Shenzhen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Speech recognition mainly refers to making the machine understand what people say, that is, in various environments, it can accurately recognize the speech content. According to the voice information, the machine can execute the intention of human. In this paper, the feature extraction algorithm of speech data set is designed. The voice data set used is thchs30, which contains 13388 voice files. The fbank feature of speech is input into the recurrent neural network for training. And, the training method is end-to-end, and the decoding result is the corresponding syllable in the dictionary. Among them, the initial and final of syllable is used as the voice label for training, and the accuracy is about 70%. After changing the mapping relationship between speech sequence and Pinyin label, about 1209 Pinyin are sorted out, and the speech features with Pinyin labels are trained. The accuracy is about 80%.

Original languageEnglish
Title of host publication2020 IEEE 5th International Conference on Signal and Image Processing, ICSIP 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages794-798
Number of pages5
ISBN (Electronic)9781728168968
DOIs
StatePublished - 23 Oct 2020
Externally publishedYes
Event5th IEEE International Conference on Signal and Image Processing, ICSIP 2020 - Virtual, Nanjing, China
Duration: 23 Oct 202025 Oct 2020

Publication series

Name2020 IEEE 5th International Conference on Signal and Image Processing, ICSIP 2020

Conference

Conference5th IEEE International Conference on Signal and Image Processing, ICSIP 2020
Country/TerritoryChina
CityVirtual, Nanjing
Period23/10/2025/10/20

Keywords

  • connectionist temporal classifier
  • deep neural network
  • recurrent neural network
  • speech recognition

Fingerprint

Dive into the research topics of 'Large Vocabulary Continuous Speech Recognition with Deep Recurrent Network'. Together they form a unique fingerprint.

Cite this