Tusom2021

Tusom2021数据集是由卡内基梅隆大学等机构创建的，包含2255条东图索语的语音转录数据，用于支持低资源和濒危语言的语音识别研究。该数据集包含丰富的语音特征，如多种辅音和元音以及三个声调，适用于测试语言无关的语音识别技术。数据集的创建过程涉及多年的转录和校正，确保了数据的质量和准确性。该数据集主要应用于自动语音识别系统的开发和评估，特别是在处理低资源和濒危语言方面具有重要价值。

The Tusom2021 dataset was created by Carnegie Mellon University and other institutions. It consists of 2255 speech transcriptions of the East Tusom language, which is developed to support speech recognition research for low-resource and endangered languages. This dataset features rich phonetic characteristics, including various consonants, vowels and three tones, making it suitable for testing language-agnostic speech recognition technologies. The construction of the dataset involved years of transcription and proofreading, ensuring its high quality and accuracy. It is mainly utilized for the development and evaluation of automatic speech recognition systems, and holds significant research value especially for processing low-resource and endangered languages.