multiple object detection

[NikitaVasilevN]

Is is possible to use this algorithm for multiple object detection? When list of label looks like this [[48], [9,9,9], [1, 1], [22]] instead of [[1,1,1], [0], [0,0,1]] (as a example). Maybe somebody have experiance with fusion bbox with yolo v8 and dino?

[tatigabru]

Do you mean for multiple classed of objects? Yes, it was applied for multiple classes on coco (80 object categories). Can you please explain a bit more, I think I did not get your question.
You merge predictions (which are confidence scores between 0 and 1), not the true labels