I have a YOLOv10 model running with C++ for inference. The output provides bounding box position information which I believe is in the format XYXY. The values I get for the first box are below:
object_conf: 0.71359
x0: -0.243503
y0: -0.257284
x1: 0.45154
y1: 0.376244
How do I convert the decimal positional information to something which I can overlay on my 640x640 images?
Thanks
@ABest2 Ultralytics only supports the Python package ultralytics and so it will be difficult to know for certain what format the coordinates are that you’re getting. Assuming the they are in fact XYXY, then the only conversion you’d need to do, is to multiply x0 and x1 by the image width and multiply y0 and y1 with the image height.
This is based on the assumption that these are image normalized coordinates. Like I mentioned, we don’t support anything that runs on C++ so it’s not certain that would be the correct method. I would suggest consulting the documentation/support for whatever framework or C++ interface your using for inference.
Thank you