Technologies on Display
F1 - Chinese Speech Recognition IC
The
main objective of this project is to develop a low-cost Cantonese and
Putonghua speaker independent speech recognition ICs using an advanced
statistical speech model and data speech library developed by the DSP group
of The Chinese University. We use a modular design approach, the basic
speech recognition engine can operate in parallel to increase the
recognition power. This basic core speech recognition engine will not only
be able to recognize one hundred isolated Chinese words in real time (less
than 1 sec), but that they will also be able to distinguish connected words
as a phrase. As far as we know, this will be the most powerful speech
recognition IC on the market. The test chip was fabricated with a 0.35
micron triple metal CMOS technology.
Applications
- Home appliances
- Smart home
- Elevator control
- Human to machine interface
Features
- Single chip solution
- Advanced HMM statistical speech model
- Large speech library
- Fast recognition time
- High recognition accuracy
- Can recognize both isolated words and connected words as a phrase
Principal Investigator
Prof. C. F. Chan
Department of Electronic Engineering
¡@ |