Research on speech recognition technology based on Wisdom Classroom

Nowadays, Wisdom classroom has many image recognition functions, but there are still large gaps and shortcomings in speech recognition. How to give real-time feedback to teachers and students on the discussion records and knowledge notes taught by teachers in class. Solving this problem requires rea...

Full description

Saved in:
Bibliographic Details
Main Authors: Luo, Kai, Zhu, Guibin, Luo, Rongjian, Zhou, Youwei, Huang, Leping, Wu, Yaoyue
Format: Conference Proceeding
Language:English
Subjects:
Automation
CNN-BiGRU
CTC-beam-search
Feature extraction
Image recognition
Network architecture
Semantics
Speech recognition
Training
Wisdom classroom
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Nowadays, Wisdom classroom has many image recognition functions, but there are still large gaps and shortcomings in speech recognition. How to give real-time feedback to teachers and students on the discussion records and knowledge notes taught by teachers in class. Solving this problem requires real-time speech recognition. The number of training samples in the system is small. Through the integration training of multiple data sets. The improved CNN-BiGRU algorithm is used to fully extract speech features. Through CTC-beam-search get the optimal recognition result.
ISSN:2689-6621
DOI:10.1109/IAEAC54830.2022.9929727