AI Sound Event Detection

Extracts Mel-frequency cepstral coefficients (MFCC) from specific sound datasets as training samples. Uses MFCC features as input and applies deep neural network algorithms to intelligently analyze and identify specific sound events like snoring or baby crying, enabling automatic alerts or responses.

Real-time response

Uses offline sound detection for quick action triggering 

High accuracy

Uses specific sound deep learning detection models 

Multiple sound types

Detects baby crying, snoring, male/female voices, etc. 

Voiceprint Recognition

Identifies speaker identity by extracting voiceprint features from speech signals. Chipintell's voiceprint recognition technology can register multiple identity authentications simultaneously, with recognition accuracy over 93%. Voiceprint recognition technology has wide applications in smart home and other fields, enabling personalized device settings, identity verification, and age restrictions.

Offline speech recognition

Offline voiceprint recognition

Offline voiceprint registration

Voiceprint Recognition

After power-on, the microphone continuously performs sound detection. When sound is detected, it calculates the feature similarity with the pre-recorded voice, and finally determines whether it matches based on a threshold to achieve speaker identification or verification. While performing voiceprint recognition, it also supports offline natural speech and fixed phrase recognition, and supports multi-intent understanding.

Chip Datasheet



Dev Board Manual



Voiceprint Recognition

Snore Detection

Automatically identifies and locates snoring sounds through steps including snoring sound signal collection, preprocessing, feature extraction, neural network model training, and real-time detection. Can be applied to products like smart mattresses, anti-snore pillows, and sleep monitors to assess sleep quality and potential respiratory issues.

Snore detection

Fast response

Snore Detection

This solution is currently only used for snore detection, not speech recognition. After power-on, it continuously performs snore detection. When snoring is detected, it outputs serial protocol and I/O signals. The serial port can be used to configure and obtain snore sensitivity parameters, or to configure whether to announce when snoring is detected.

Chip Datasheet



Module Dadasheet



Dev Board Manual



Snore Detection

Cry Detection

Automatically identifies and locates baby crying sounds through steps including sound signal collection, preprocessing, feature extraction, neural network model training, and real-time detection. Can be applied to products like baby monitors, cribs, and baby toys.

Cry detection

Fast response

Cry Detection

This solution is currently only used for cry detection, not speech recognition. After power-on, it continuously performs cry detection. When crying is detected, it outputs serial protocol and I/O signals. The serial port can be used to configure and obtain cry sensitivity parameters, or to configure whether to announce when crying is detected.

Chip Datasheet



Module Dadasheet



Dev Board Manual



Cry Detection

Male/Female Voice Detection

Builds voiceprint models by extracting feature parameters from male and female voice signals to provide personalized services based on gender differences. Can be applied to smart toilets, game character props, etc.

Male/Female voice detection

Fast response

Male/Female Voice Detection

After power-on, the microphone continuously performs sound detection. When sound is detected, it calculates the feature similarity with the pre-recorded voice, and finally determines whether it matches based on a threshold to achieve speaker identification or verification. While performing male/female voice detection, it also supports offline natural speech and fixed phrase recognition, and supports multi-intent understanding.

Chip Datasheet



Module Dadasheet



Dev Board Manual



Male/Female Voice Detection