To help Malia communicate more accurately and more fully, the speech translator team developed a software to learn from and adapt to her voice, understand what she is saying and translate it to others around her. The team made up of a combination of Berkeley students and Googlers, developed an algorithm that takes in samples of Malia’s voice and trains itself to pick up on the nuances of her speech.
The product needs to be portable and carried on her wheelchair. As a first stage, any laptop or tablet is good enough. The product needs to require no actions by the user (Malia), except her normal speech. The operator (the person listening to her) can operate the user-interface but it should be a simple one.
Malia is an 11 year old. Malia suffers from Cerebral palsy and as a result her speech is very hard to understand by her environment. Her close family can understand most of what she says, but it is hard for others to understand her. It causes her great frustration when people don’t understand her and it discourages her speaking.
The build files for Speech Translator are free to download from this page - open source, shared by TOM: Tikkun Olam Makers.
Browse all TOM projects · Search projects · TOM communities · Makeathons and events