Getting x, y, height, width of object in rendered image

Hi there,

I am using unreal engine to test the accuracy of object detection algorithms (YOLO, SSD). Currently, we render an image and then use labelImg to label objects and draw a box around them. For example, it might be an image with a car in the centre. We draw a box using labelimg and get a file that describes the x and y location, plus the height and width of the box.

Might there be any way to use unreal engine to generate these box labels also?