The Thai Elderly Speech dataset by Data Wow and VISAI Version 1 dataset aims at advancing Automatic Speech Recognition (ASR) technology specifically for the elderly population. Researchers can use this dataset to advance ASR technology for healthcare and smart home applications. The dataset consists of 19,200 audio files, totaling 17 hours and 11 minutes of recorded speech. The files are divided into 2 categories: Healthcare (relating to medical issues and services in 30 medical categories) and Smart Home (relating to smart home devices in 7 household contexts). The dataset contains 5,156 unique sentences spoken by 32 seniors (10 males and 22 females), aged 57-60 years old (average age of 63 years).
Dataloader name:
thai_elderly_speech/thai_elderly_speech.py
DataCatalogue: http://seacrowd.github.io/seacrowd-catalogue/card.html?thai_elderly_speech