Using Character-Level Sequence-to-Sequence Model for Word Level Text Generation to Enhance Arabic Speech Recognition

Using Character-Level Sequence-to-Sequence Model for Word Level Text Generation to Enhance Arabic Speech Recognition

Owing to the linguistic richness of the Arabic language, which contains more than 6000 roots, building a reliable Arabic language model for Arabic speech recognition systems faces many challenges. This paper introduces a language model free Arabic automatic speech recognition system for Modern Stand...

Full description

Bibliographic Details
Main Authors: Mona A. Azim, Wedad Hussein, Nagwa L. Badr
Format: Article
Language:English
Published: IEEE 2023-01-01
Series:IEEE Access
Subjects:
Arabic speech recognition
CNN
CTC
RNN
Online Access:https://ieeexplore.ieee.org/document/10209196/

Internet

https://ieeexplore.ieee.org/document/10209196/

Similar Items