ChildMandarin
收藏资源简介:
ChildMandarin是一个专为3至5岁儿童设计的普通话语音数据集,由南开大学计算机科学学院创建。该数据集包含41.25小时的语音数据,来自397名儿童,覆盖中国22个省级行政区,性别分布均衡。数据集通过精心设计的手动转录,确保了高质量的语音标注。创建过程中,数据集考虑了儿童语音的独特性,包括发音、语调和语速的差异。该数据集主要应用于教育技术和儿童与计算机交互领域,旨在解决现有自动语音识别系统在儿童语音识别上的不足。
ChildMandarin is a Mandarin speech dataset specifically designed for children aged 3 to 5 years old, developed by the School of Computer Science at Nankai University. This dataset comprises 41.25 hours of speech data collected from 397 children across 22 provincial-level administrative regions in China, with a well-balanced gender distribution. It employs rigorously designed manual transcription to guarantee high-quality speech annotations. During its development, the dataset fully accounts for the unique characteristics of children’s speech, including variations in pronunciation, intonation and speech rate. It is primarily utilized in educational technology and the field of human-computer interaction for children, with the objective of addressing the limitations of current automatic speech recognition systems in children’s speech recognition tasks.

- 1ChildMandarin: A Comprehensive Mandarin Speech Dataset for Young Children Aged 3-5南开大学计算机科学学院 · 2024年



