AI Chat AI Image Generator AI Video Text to Speech

The CUHK-TUDELFT System for The SLT 2021 Children Speech Recognition Challenge

11/12/2020

∙

by Si-Ioi Ng, et al.

∙

∙

This technical report describes our submission to the 2021 SLT Children Speech Recognition Challenge (CSRC) Track 1. Our approach combines the use of a joint CTC-attention end-to-end (E2E) speech recognition framework, transfer learning, data augmentation and development of various language models. Procedures of data pre-processing, the background and the course of system development are described. The analysis of the experiment results, as well as the comparison between the E2E and DNN-HMM hybrid system are discussed in detail. Our system achieved a character error rate (CER) of 20.1 designated test set, and 23.6 at 10-th overall.

Si-Ioi Ng
8 publications
Wei Liu
370 publications
Zhiyuan Peng
11 publications
Siyuan Feng
29 publications
Hing Pang Huang
2 publications
Odette Scharenborg
26 publications
Tan Lee
41 publications

page 1

page 2

page 3

page 4

research

∙ 11/09/2020

Data Augmentation For Children's Speech Recognition – The "Ethiopian" System For The SLT 2021 Children Speech Recognition Challenge

This paper presents the "Ethiopian" system for the SLT 2021 Children Spe...

0 Guoguo Chen, et al. ∙

research

∙ 02/19/2021

End-to-End Neural Systems for Automatic Children Speech Recognition: An Empirical Study

A key desiderata for inclusive and accessible speech recognition technol...

0 Prashanth Gurunath Shivakumar, et al. ∙

research

∙ 10/02/2021

Significance of Data Augmentation for Improving Cleft Lip and Palate Speech Recognition

The automatic recognition of pathological speech, particularly from chil...

0 Protima Nomo Sudro, et al. ∙

research

∙ 05/25/2016

On model architecture for a children's speech recognition interactive dialog system

This report presents a general model of the architecture of information ...

0 Radoslava Kraleva, et al. ∙

research

∙ 06/14/2023

Research on an improved Conformer end-to-end Speech Recognition Model with R-Drop Structure

To address the issue of poor generalization ability in end-to-end speech...

0 Weidong Ji, et al. ∙

research

∙ 05/05/2020

End-to-end Whispered Speech Recognition with Frequency-weighted Approaches and Layer-wise Transfer Learning

Whispering is an important mode of human speech, but no end-to-end recog...

0 Heng-Jui Chang, et al. ∙

research

∙ 10/08/2020

Population Based Training for Data Augmentation and Regularization in Speech Recognition

Varying data augmentation policies and regularization over the course of...

0 Daniel Haziza, et al. ∙