Hello, im about to buy the openMV H7 plus cam and I am worried if I will be able to use Tenserflow Lite models. Also if i do use a tenserflow lite model on the cam to detect green,red,silver balls,black balls and blue cubes will i be expecting a massive reduction in fps and resolution?
Hi, you can run tensorflow lite models on the system trained by edge impulse. The fomo model can hit 30 FPS.
However, if you are doing color tracking I recommend your find_blobs() it will be faster than tensorflow for something this simple.
thanks i’ll check that function out