It realizes isolated word recognition (phrase recognition) for speech recognition.
When you select one phrase from the pre-registered phrases and input it via voice, the recognition engine (voice recognition LSI) compares it with the group of phrases in the database (training data) and calculates the index assigned to the closest phrase along with its likelihood value. You can choose one of three types of host serial interfaces for the connection with the host.
Inquire About This Product
basic information
【Features】 ○ Supports unspecified speakers (no individual training required) ○ Achieves high noise robustness (recognition rate of over 98% at S/N ratio of 20dB) ○ Automatic recognition start function *a. Patent pending ○ Utilizes a dedicated hardware engine to achieve functionality at relatively low clock frequencies ○ Function to eliminate false responses to inputs other than registered words *a (misrecognition rejection function) ○ Equipped with three types of host serial interfaces (SPI/I2C/UART) ○ Collaborative research with universities to further advance speech recognition technology ● For other functions and details, please contact us.
Price information
Please feel free to contact us.
Delivery Time
※Please feel free to contact us.
Applications/Examples of results
【Usage】 ○SPI: Maximum 5Mbps ○I2C: Maximum 400kbps (high-speed mode) byte access ○UART: Character transfer at 9600/19200/38200/57600bps ●For more details, please refer to the catalog or contact us.
catalog（1）
Download All Catalogs
Company information
We provide comprehensive support from project and specification decision-making to functional design, theoretical design, layout design, and process design. We have equipped ourselves with the latest design tools and take pride in our environment and track record that can certainly meet advanced needs.
