Melodic is an integrated system that provides professional speech therapy for children with hearing impairment via a multisensory edge device and web applications. Training activities provided by Melodic cover breath training, mouth shape training, and sound training. Supported by Microsoft Azure, we developed a highly scalable system that utilizes cognitive analysis for mouth movement recognition, storage and analytics for behavior data, and power BI and app service for web applications. Around 1.5 billion people live with hearing loss in the world, and 430 million of them can speak after speech rehabilitation, especially prelingual deaf children. In contrast, speech pathologist resource is limited, the number of which in the US is only 120,094. Besides, traditional speech therapy costs about $24,000 per year, which can be a high expenditure for most families. To ameliorate the current situation, Melodic provides customized services to both children and pathologists. With physical interaction and big data analysis, it ensures the professional degree of therapy care while keeping a low cost compared to existing products. Taking a long-term view, we envision Melodic growing as a platform based on Azure. By providing hardware and software interfaces, Melodic can be empowered by more customized training activities developed by the community.
We are students from MSTI program at University of Washington. Our group members are Xiaoyang Qiu, Jinyao Ouyang, Shuoxuan Wang.