Research Article
Speech Processing: a literature review
@INPROCEEDINGS{10.4108/eai.18-12-2023.2348132, author={Go Issa Traore and Borlli Michel Jonas Some}, title={Speech Processing: a literature review}, proceedings={Proceedings of the 6th Computer Science Research Days, JRI 2023, 18-20 December 2023, Ouagadougou, Burkina Faso}, publisher={EAI}, proceedings_a={JRI}, year={2024}, month={6}, keywords={speech processing machine learning speech features extractor speech database}, doi={10.4108/eai.18-12-2023.2348132} }- Go Issa Traore
Borlli Michel Jonas Some
Year: 2024
Speech Processing: a literature review
JRI
EAI
DOI: 10.4108/eai.18-12-2023.2348132
Abstract
In this paper, we have focused our research on the state of the knowledge on speech processing and the research perspectives that exist in this domain. This research was conducted on several digital libraries such as IEEE Xplore, ScienceDirect, arXiv, Springer Link, Papers With Code etc. The research focused on the types of speech classification, the techniques used to extract speech features, the Machine Learning (ML) techniques used and the speech data sources available. We found that studies focused mainly on emotion recognition, dialect identification in speech and speaker recognition. Mel Frequency Cepstral Coefficents (MFCC) is the main and most widely used for speech feature extraction. Neural networks dominate as ML techniques for speech classification. Speech databases available have been built in different contexts. Each database is specific to a given language, mainly English, German, Arabic, Chinese and French. There are almost no speech databases for low-resource languages, particularly african languages.