Scholarlink

This project aims to analyse one of the most practical and easily produced biometric features: speech, using intelligent approaches. Technological developments over the last twenty years have made it possible to convert audio speech into text, in particular to help the hearing impaired. However, our work is not limited to simply transcribing speech; it also includes assigning the speaker and author after transcription. The classification models used in this task include MLP, SVM, KNN, CNN and Transformer.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automatic Speech Transcription for Author and Speaker Attribution Using Deep Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • S. Bourib,
  • Y. Aichouba,
  • N. Aguini,
  • A. Malaoui

摘要

This project aims to analyse one of the most practical and easily produced biometric features: speech, using intelligent approaches. Technological developments over the last twenty years have made it possible to convert audio speech into text, in particular to help the hearing impaired. However, our work is not limited to simply transcribing speech; it also includes assigning the speaker and author after transcription. The classification models used in this task include MLP, SVM, KNN, CNN and Transformer.