Silent speech device developed by Cornell Students
About this video
Cheng Zhang (CIS) and graduate student Ruidong Zhang have developed SpeeChin, the first necklace-based silent-speech recognition technology that can recognize 54 English and 44 Chinese silent speech commands. It uses a customized neck-mounted IR camera to capture the images of the neck and face from under the chin.
Related article summary
Cornell researchers developed SpeeChin, a necklace-worn infrared camera that captures chin and neck movements to recognize silent, whispered, or mouthed speech. It distinguishes 54 English phrases and 44 Mandarin words, achieving 90.5% accuracy for English and 91.6% for Chinese in tests with 20 people; accuracy falls when users move. The device aims to enable discreet voice commands in quiet or socially sensitive settings but requires further reliability improvements before commercial availability.