AI acceleration gesture recognition experience based on efr32mg24

        With AI In depth , In the early AI They can only run on high-performance chips , stay MCU Basically difficult , With the development of these years, gradually appeared MCU AI programme , Many manufacturers have also launched targeted MCU Development plan of , among edgeimpulse and sensiml Yes MCU Developers are the fastest to get started , It is very suitable for the application scenario of classification feature recognition .

        AI The process is basically the same ： Data collection and annotation （ It is divided into training set and test set ）, The design model （ Generally, there are direct modifications available ,）, Training models , test model , Deploy to the actual chip for actual testing .

Data collection annotation

Model definition

model training

 

Model test

 

Model export after training

Deploy to the actual prototype for actual testing

 

MG24 AI

 

It's OK to try it as a whole , With this, you can quickly do a simple speech recognition, etc AI Recognition application .