Thanks for the pointers! Yes we are expecting an uphill struggle trying to get it to work haha. The output of a computer vision detection model is quite monsterous (especially since I do not understand all of it!). We will probably have to rely on a lot of code-borrowing from the yolo repositories haha.