概览
描述
Cyberon Neural Network model-based Speaker Verification (CNSV) is a speaker identification solution on an MCU platform. CNSV has no constraint on speech content with low recognition error rates and delivers high robustness by the Neural Networks model. It supports multiple speaker identification depending on system resources. Available now for RA8M1 and RA6E1 Renesas MCUs.
特性
- Utilizing Arm® CMSIS-NN to enhance computation efficiency on Cortex®-M MCU
- Enables identification among multiple registered speakers
- Text-independent solution
- No network connection is needed
- Low recognition error rates and high robustness by Neural Networks model
- Small memory footprint and with multiple options (Flash: 715KB ~ 900KB, RAM: 200KB ~ 245KB)
- Higher recognition performance with optimization by Arm Helium™ on RA8
- Low power and high efficiency MCU with strong security function
- Combination with Cyberon Dspotter or DspotterNLU provides more comprehensive voice solution
产品对比
应用
视频和培训
The smart voice, no cloud solution is Renesas’ AI Voice Combo Solution powered by the VK-RA8M1 kit. It delivers advanced voice control without relying on cloud connectivity. The solution features APF for noise isolation, NLU for natural speech recognition, CNSV for speaker verification, and VAS for anti-spoofing. These capabilities enable secure and intuitive user interaction.
Related Resources