Changing the shape of inference images

It was written on the website that we can scale up the mages for inferencing even if we have trained it on a smaller patch of images[fomo]( How do we do that on my local model in c++ ( downloaded as c++ library)?

You can download the Arduino Library, open one of the associated examples for say the Portenta H7 in the Arduino IDE and look for bool ei_camera_capture() wherein image re-sizing takes place.