I’m am aware of three alternative approaches, namely i) instance segmentation, ii) semantic segmentation and iii) panoptic segmentation and my understanding is their difference lies in considering just objects’ classes (semantic), just objects’ instances (instance) or both (panoptic).
My question is: if I want to count objects in an image and I want to provide also a justification of the final count as a segmentation mask, what of these 3 approaches may best fit the purpose?
Thanks in advance 