Noise Reduction

The dual microphone noise reduction technology is based on the latest neural network processing speech algorithm, which effectively solves the problems of speech noise pollution and indoor reverberation interference, improving the intelligibility and recognition rate of speech signals. After receiving two audio inputs from the dual microphone, the signal source and noise source are identified through speech enhancement algorithms, effectively suppressing noise and removing certain reverberation and echo interference, further improving the effectiveness and quality of sound collection, providing effective guarantees for subsequent activation and command word recognition.

Face Recognition

M4V3 is a facial recognition module solution customized and developed for the intelligent door lock industry, compatible with the vast majority of lock plate solutions on the market. The aim is to empower AI in the intelligent door lock industry, promote industry upgrading, and create safer and more reliable intelligent door lock products. The M4V3 adopts facial recognition technology, equipped with biomimetic binocular facial recognition algorithms, and has been specially optimized for intelligent door lock application scenarios, achieving industry-leading security while also providing a good user experience, such as fast recognition speed, low power consumption, strong environmental adaptability (fearless of various complex lighting environments), and wide height coverage.

Offline Voice

1. Pure offline solution, voice control can be achieved even when not connected to the network 2. Adopting FreeRTOS, the system starts quickly after power on 3. Meet the voice interaction needs of home appliances: a. Dual microphone front-end processing algorithm, meeting most environmental noise reduction needs b. New generation neural network activation and recognition technology c. Support for built-in audio playback 4. Low error activation rate superior to industry level 5. High recognition rate for Mandarin with dialect accents 6. Module communication interfaces are UART and USB. 7. Module size: 38mm * 46.1mm

On&Off-line Voice

1. Offline+online dual mode, with voice interaction available whether connected or not 2. Adopting FreeRTOS, the system starts quickly after power on to meet the voice interaction needs of home appliances: a. Dual microphone neural network front-end processing algorithm, meeting most environmental noise reduction needs b. Echo cancellation, support for broadcast interruption c. New generation neural network activation and recognition technology d. Supports built-in audio playback 3. Low error activation rate superior to industry level 4. High recognition rate for Mandarin with dialect accents 5. Cloud access to iFLYOS/AIUI voice operating system a. Support dialect recognition and multilingual recognition b. The naturalness score of speech synthesis leads the world c. Cloud semantic understanding only takes 100ms 6. Module size: 38.1mm * 60.2mm