AI MODEL DETECTS NOISES BY CONVERTING SOUNDS INTO SPECTROGRAM IMAGES.
Sound Model was born to find noises during the testing and validation of braking systems, but it can learn from your experts how to detect other unpleasant sounds on any product/good during its operation.
Detection is done by cleverly converting the sound of a microphone into images suitable to feed our AI, then detecting sounds as if they where objects in the images.
Every sound/noise in fact leaves a specific fingerprint in a so-called “spectrogram” image.
Every sound/noise in fact leaves a specific fingerprint in a so-called “spectrogram” image.
ARCHITECTURE
Sound Model uses a highly specialized Deep Neural Network to learn new sounds with a low number of examples.
The AI algorithm runs into a server, on-edge or in-cloud, receiving sounds from microphones and giving back the list of unwanted noises found in there.
An optional User Interface can be designed upon request to match the requirements of the business.
The AI algorithm runs into a server, on-edge or in-cloud, receiving sounds from microphones and giving back the list of unwanted noises found in there.
An optional User Interface can be designed upon request to match the requirements of the business.
