一项用于自动语音识别的概念验证研究,用于从高科技 AAC 系统中转录 AAC 扬声器的语音
Szu-Han Kay Chen1, Conner Saeli2, Gang Hu2
1Department of Communication Sciences and Disorders, The University of New Hampshire, Durham, New Hampshire, USA.
Assistive technology : the official journal of RESNA
|September 25, 2023
概括
一个新的增强和替代通信-自动语音识别 (AAC-ASR) 模型准确地转录来自高科技 AAC 系统的语音. 这项技术有助于对使用AAC设备的个人进行语言样本分析.
科学领域:
- 语音和语言病理学 语言病理学
- 计算机科学 计算机科学
- 辅助技术 辅助技术 辅助技术
背景情况:
- 自动语音识别 (ASR) 用于非典型的语音和语言样本的转录.
- 在高科技增强型和替代型通信 (AAC) 系统中使用ASR仍然未被探索.
- 语言样本分析在沟通科学和障碍学中至关重要.
研究的目的:
- 研究使用AAC-ASR模型来转录来自高科技AAC系统的语言样本的可行性.
- 将开发的AAC-ASR模型的转录精度与已建立的ASR模型 (CMU Sphinx,谷歌语音到文本) 进行比较.
主要方法:
- 开发了一个概念验证的AAC-ASR模型来转录模拟的AAC扬声器语言样本.
- 使用测试文件评估了AAC-ASR模型的Word错误率 (WER).
- 将AAC-ASR模型的WER与CMU Sphinx和谷歌语音到文本进行了比较.
主要成果:
- 开发的 AAC-ASR 模型实现了 28.6% 的 Word 错误率 (WER).
- 这在测试的AAC语音样本上显著超过了CMU Sphinx (70.7% WER) 和谷歌语音到文本 (86.2% WER).
- 证明了对高科技 AAC 模拟语言样本的自动转录的可行性.
结论:
- 在AAC-ASR模型显示承诺自动转录高科技AAC生成的语言样本.
- 这项技术可以支持通信科学中的语言样本分析.
- 未来的研究应该专注于各种AAC语音数据集和个人用户语音模式的完善.
相关概念视频
01:17Transcription
Transcription is the synthesis of RNA from a DNA sequence by RNA polymerase. It is the first step in producing a protein from a gene sequence. Additionally, many other proteins and regulatory sequences are involved in correctly synthesizing messenger RNA (mRNA). Transcriptional regulation is responsible for the differentiation of different types of cells and often for the proper cellular response to environmental signals.
Transcription Can Produce Different Kinds of RNA Molecules
In eukaryotes,...
Transcription Can Produce Different Kinds of RNA Molecules
In eukaryotes,...
01:28Automatic Processing and Automatic Social Behavior
Automatic processing refers to the cognitive operations that occur without conscious intent or awareness, playing a fundamental role in shaping social cognition and behavior. These processes enable individuals to navigate complex social environments efficiently by relying on mental shortcuts and pre-existing knowledge structures known as schemas. One of the most influential mechanisms underlying automatic processing is priming, which subtly activates mental representations through exposure to...