Spoken Language assessment on Mobile Devices (Ongoing)
PI: Prof. Preeti Rao, Department of Electrical Engineering
Fellow: Shreeharsha B S, Vempalli Naga Sai Saketh (RA) (2020-2022)
A shortage of teachers and absence of speaking opportunities outside school contribute to a lack of practice in spoken English for children in rural schools. The project seeks to alleviate the problem by targeting an affordable and scalable technology solution that can help struggling learners with automatic means to obtain immediate feedback on pronunciation and fluency. Children are encouraged to read stories aloud on a mobile device using a headset mic in video karaoke mode where words are highlighted at normal reading pace together with the option of a guiding narrator voice. The child’s recorded reading is processed using speech recognition technology to provide feedback on reading accuracy and fluency. The automatic assessment system is developed using a large dataset of oral reading recordings from the target group of children of various proficiency levels as annotated by human experts.
