Called MOVI (My Own Voice Interface), it is the work of Bertrand Irissou and Gerald Friedland founders of Audeme and the project will launch on the Kickstarter crowd-funding website at 8am Wednesday 1 July. (11:59pm PDT)
MOVI is cloudless with no Internet connection required, and has everything built into the board, including an English dictionary for full sentence recognition, and is ready to go right out of the box.
The design aim was to build a self-contained speech recognition system which does not rely on a connection to the internet.
They believe they have achieved this with an ARM processor-based board capable of full sentence recognition and automatic phonetisation.
The developers write:
“It includes an English phonetisation dictionary of over 2GB in order to be able to recognize virtually any English sentence. The power requirements are extremely low, about 2.5W, making it suitable for potential battery operation”.
Power consumption is less than 3W.
Irissou was the founder of Asic Advantage which was acquired in 2011 by Microsemi.
Friedland has contributed to the academic and open-source communities through his research at UC Berkeley, where he is currently director of the audio and multimedia research at the International Computer Science Institute.
More on the Audeme Facebook page.
from News http://ift.tt/1LAp3SO
via Yuichun
沒有留言:
張貼留言