Search research articles

关于 JoVE

概览领导团队博客 JoVE 帮助中心

作者

出版流程编辑委员会范围与政策同行评审常见问题投稿

图书馆员

用户评价订阅访问资源图书馆顾问委员会常见问题

研究

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments 存档

教育

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual 教师资源中心教师网站

使用条款与条件

相关概念视频

Improving Translational Accuracy

Improving Translational Accuracy

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序

Same author

Effects of mindfulness meditation on attention processes in people with aphasia.

International journal of speech-language pathology·2026

Same author

Towards AI-Based Strep Throat Detection and Interpretation for Remote Australian Indigenous Communities.

Sensors (Basel, Switzerland)·2025

Same author

Discriminating Children with Speech Sound Disorders from Children with Typically Developing Speech Using the Motor Speech Hierarchy Probe Words: A Preliminary Analysis of Mandibular Control.

Diagnostics (Basel, Switzerland)·2025

Same author

Using Inertial Measurement Units and Machine Learning to Classify Body Positions of Adults in a Hospital Bed.

Sensors (Basel, Switzerland)·2025

Same author

Facial Movements Extracted from Video for the Kinematic Classification of Speech.

Sensors (Basel, Switzerland)·2024

Same author

Underwater Acoustic Orthogonal Frequency-Division Multiplexing Communication Using Deep Neural Network-Based Receiver: River Trial Results.

Sensors (Basel, Switzerland)·2024

Search research articles

相关实验视频

Updated: Jul 7, 2025

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Published on: September 27, 2024

改进文本独立的强制对齐,以支持语音语言病理学家使用语音转录.

Ying Li¹, Bryce Johannas Wohlan¹, Duc-Son Pham¹

¹School of EECMS, Curtin University, Bentley, WA 6102, Australia.

Sensors (Basel, Switzerland)

|December 23, 2023

概括

这项研究引入了一种自动语音转录模型,用于诊断语音障碍 (SSD). 这种新的方法提高了准确性,减少了语音分析中的偏见,有利于临床应用.

关键词:

强制对齐的强制对齐方式音名细分语音细分语音细分声学障碍声学障碍声学障碍声学障碍语音声音障碍语音声音障碍语音治疗治疗语音治疗 wav2vec 2.0 的使用情况.

更多相关视频

Ultrasound Images of the Tongue: A Tutorial for Assessment and Remediation of Speech Sound Errors

Ultrasound Images of the Tongue: A Tutorial for Assessment and Remediation of Speech Sound Errors

Published on: January 3, 2017

A Protocol for Comprehensive Assessment of Bulbar Dysfunction in Amyotrophic Lateral Sclerosis ALS

A Protocol for Comprehensive Assessment of Bulbar Dysfunction in Amyotrophic Lateral Sclerosis ALS

Published on: February 21, 2011

相关实验视频

Last Updated: Jul 7, 2025

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Published on: September 27, 2024

Ultrasound Images of the Tongue: A Tutorial for Assessment and Remediation of Speech Sound Errors

Ultrasound Images of the Tongue: A Tutorial for Assessment and Remediation of Speech Sound Errors

Published on: January 3, 2017

A Protocol for Comprehensive Assessment of Bulbar Dysfunction in Amyotrophic Lateral Sclerosis ALS

A Protocol for Comprehensive Assessment of Bulbar Dysfunction in Amyotrophic Lateral Sclerosis ALS

Published on: February 21, 2011

科学领域:

语言病理学语音病理学
计算语言学计算语言学
机器学习机器学习

背景情况:

语音转录对于诊断语音音障碍 (SSD) 是至关重要的.
当前的强制对齐 (FA) 工具通常需要手动转录,并且容易产生偏差.
现有的FA工具的局限性阻碍了有效的语音模式诊断和分析.

研究的目的:

开发一种新的,文本独立的强制对齐模型,用于自动语音转录.
解决手动转录和感知偏差在SSD诊断中的局限性.
提高语音障碍评估的客观性和效率.

主要方法:

使用预训练的 wav2vec 2.0 模型进行自动语音分割和识别.
使用无监督细分工具 (UnsupSeg) 准确识别语音边界.
实现了近邻分类,连接式时间分类 (CTC) 和后处理以增强细分.

主要成果:

该模型实现了竞争性表现,在TIMIT数据集 (普通扬声器) 上,其和平均得分为76.88%.
首次,该模型在TORGO数据集 (SSD扬声器) 上进行了评估,达到70.31%的和平均得分.
在有言语声音障碍的扬声器上表现出有效的性能,这是一个关键的进步.

结论:

开发的模型在对SSD的客观和不那么偏见的评估方面取得了重大进展.
它与SSD扬声器的有效性为语音病理学开辟了新的研究和临床应用.
这种自动化方法有可能彻底改变语言障碍的诊断和治疗.