Voiceitt
Speech recognition learns one user's non-standard speech and converts it into text or a comprehensible computer voice.
About
Voiceitt is designed for people who are not understood by ordinary speech recognition due to, for example, a neurological condition or speech impairment. The user records training sentences, after which a personal model learns recurring sounds. The software can then display speech as text or play it back through a computer voice. Results vary per person and continued training is required.
AI factor
General speech models are mainly trained on standard speech and lack individual patterns. A personal model instead learns from a single user, thereby supporting communication without taking over the user's own message.
Sources
Inspiration sourced from AI For Good. Not an Arnaldus customer case.