I’m using the Vision framework to run an object detection model on iOS but the predicted bounding boxes are not being drawn in the right place. Help! This is a very common problem and it happens because there are several different coordinate systems that you have to translate between.