Question from Predictive Maintenance Webinar 29th of April: Does the data pre-processing for ultrasound include aother techniques like Mel frequency cepstral coefficient or Linear frequency cesptral coefficient?

[Vladimir JANOUSEK]

The STWIN ultrasound example is more based on traditional FFT of the audio signal. MEL coefficients are indeed method commonly used in audio applications and stwin example could be extended for this. As additional information, you can also have look on another development kit sensortile.box "Baby Monitor" demo, this is based on analyzing MEL coefficients based on audio captured by microphone and AI type of algorithm running on STM32.