Details
Participants: 1 Student
Keywords: Audio Analysis, Sentiment Analysis, Deep Learning, App Programming
Task Description & Requirements
Implement a mobile app + server backend for the following purpose: Record live audio, recognize the voice of the owner, perform speech-to-text and then, investigate whether this speech contains elements of aggression, racism and/or sexism + related issues. Potential add-ons include: (1) a coaching function for rhetorics (e.g. suggestion of synonyms, etc.), (2) automated translation to other langauges, (3) a trainer function for improving your (singing) voice). For the base application, recording + speaker recognition should be performed on the device, while the latter steps should be performed server-based using state of the art AI libraries. Requires excellent data science skills + good software engineering skills. If performed as a diploma thesis, must include a user-based evaluation with a selected group of participants.