Amharic Speech Recognition Based on Phoneme Context-Sensitivity Analysis and Error Diagnostics

Amharic, the official language of Ethiopia, presents distinct challenges for automatic speech recognition due to its complex phonological system. Features such as ejective consonants, systematic gemination, and context-dependent vowel realizations create acoustic variability that conventional HMM an...

Popoln opis

Bibliografske podrobnosti

Principais autores:	Eyasu Demissie Yenieneh, Surafiel Habib Asefa, Yaregal Assabie
Format:	Artigo
Jezik:	Inglês
Izdano:	IEEE 2026-01-01
Serija:	IEEE Access
Teme:	Automatic speech recognition amharic speech recognition context sensitivity low-resource languages transformer Wav2vec
Online dostop:	https://ieeexplore.ieee.org/document/11435349/
Oznake:	Označite Brez oznak, prvi označite!