On this page in the docs, the following is written:
The detected bounding boxes are converted to segmentation masks by masking all pixels within boxes for performance considerations.
But in the object detection package for YOLOv8, this functionality doesn’t seem to be implemented.
What I am trying to do is to compute a segmentation mask for the whole bounding box region, and then pass this to nvBLOX.
Is there a reason why the aforementioned functionality has not been implemented yet? And is there a timeline for when this may happen?
Edit: Could someone maybe clear up how exactly the Image message should be populated for nvBLOX to successfully handle this data? Should the region of the mask be populated with int values 255, or ints corresponding to the various calls id’s?
