Another Step Forward in Automatic Speechreading

Bit by bit, computer vision analysis is advancing to the point of offering rapid, accurate speechreading. Cornell’s LipNet uses sentence-length, natural utterances to improve its accuracy, which at 93%, surpasses skilled humans. Just a prototype for now, we hope to see camera software like this widely available soon. Not just for individual communication, automatic speechreading offers better instant transcripts and captions. Better learn to cover your mouth when you mutter.

This entry was posted in Communication, Computer vision, Deaf, Hard of hearing, Just a prototype so far, Redundant output, Speech-to-text. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *