Use a ROI instead of the whole image

Project Discussion
1

Hi, I’m a beginner in OpenMV. I have an OpenMV H7 Plus, and I’ve been trying to use ImageNetV2 to train a model and test it on the OpenMV Cam.

I believe my model is working on the OpenMV Cam, but my next goal is to use a Region of Interest (ROI) to make predictions instead of using the entire image captured by the camera. For instance, I would like to define a rectangle that encompasses the area of the image I want to classify with my model. It would be great if I could pass this rectangle to the model as a ROI instead of the entire image.

Thank you

2

Use the roi=(x, y, w, h) feature of the tf module for the functions inside of it?

It’s been there and works.