IIITH-TISA
收藏资源简介:
IIITH-TISA数据集是由国际信息技术学院 - 海得拉巴与印度口吃协会(TISA)合作创建的,是首个印度英语口吃语料库。该数据集包含10小时的口吃者语音数据,经过精心标注,涵盖五种口吃类型和无口吃段落,共计3251个音频片段。数据集的创建过程严格遵循SEP-28k协议,确保了数据的高质量和一致性。该数据集主要用于语音助手的改进和口吃早期检测,旨在解决语音助手过早中断口吃者语音的问题,并为儿童口吃的早期诊断提供支持。
The IIITH-TISA Dataset was co-developed by the International Institute of Information Technology, Hyderabad and the Indian Stuttering Association (TISA), and it is the first-ever Indian English stuttering corpus. This dataset contains 10 hours of speech data from individuals who stutter, with meticulously annotated content covering five types of stuttering and non-stuttering speech segments, totaling 3251 audio clips. The dataset was constructed in strict compliance with the SEP-28k protocol to ensure high data quality and consistency. It is primarily used for improving voice assistants and early stuttering detection, with the dual objectives of addressing the issue where voice assistants prematurely interrupt the speech of people who stutter, and providing support for the early diagnosis of stuttering in children.

- 1Typical vs. Atypical Disfluency Classification: Introducing the IIITH-TISA Corpus and Temporal Context-Based Feature Representations国际信息技术学院 - 海得拉巴 · 2024年



