Nasa researchers are working on a way to get speech support
without talking. Using subvocalization
sensors that measure neuro-electric signals, they are developing
communication protocols that will be useful in situations where
normal speech might be hard to accomplish.
"In space, no one can hear you
scream. Use a cell phone on a crowded commuter train and everyone
can. Charles Jorgensen is working to solve
both problems, using an uncanny technology called subvocal speech
recognition......"..more background there
Ames
Extension of the Human Senses website:
"Benefit
The overarching goal of this project is to increase safety,
efficiency, and reliability of human-machine systems by applying
advanced sensing, human modeling, and adaptive automation
technologies. We aim to develop multimodal interfaces, which
allow humans to interact with automated systems in a natural
manner more closely approximating human communication, and to
provide automated systems with greater insights into user
intentions. Because the technology will not require traditional
keyboard or joystick-like controls, it will also provide
redundant channels for control during emergencies. Another
benefit of the sensing technology will be the provision of a
neurobehavioral data stream for automated crew health
monitoring."...more neat stuff there
Sub-auditory
speech recognition (.doc)
There was some "speech recognition" software a while back that did lip reading and had a much higher accuracy than audio-domain speech recognition at a lower CPU cost. That seems like a gimme with camera phones. Is anybody working on it?
Sub Vocal Speech
Nasa researchers are working on a way to get speech support without talking. Using subvocalization sensors that measure neuro-electric signals, they are developing communication protocols that will be useful in situations where normal speech might be hard to accomplish.
"In space, no one can hear you scream. Use a cell phone on a crowded commuter train and everyone can.
Charles Jorgensen is working to solve both problems, using an uncanny technology called subvocal speech recognition......"..more background there
Ames Extension of the Human Senses website:
"Benefit
The overarching goal of this project is to increase safety, efficiency, and reliability of human-machine systems by applying advanced sensing, human modeling, and adaptive automation technologies. We aim to develop multimodal interfaces, which allow humans to interact with automated systems in a natural manner more closely approximating human communication, and to provide automated systems with greater insights into user intentions. Because the technology will not require traditional keyboard or joystick-like controls, it will also provide redundant channels for control during emergencies. Another benefit of the sensing technology will be the provision of a neurobehavioral data stream for automated crew health monitoring."...more neat stuff there
Sub-auditory speech recognition (.doc)