Instance Segmentation: Counting Cars from Public CCTV with YOLOv8 seg

The 0.604 versus 0.668 gap is a good quantitative version of the point, and it is the shape you would predict: boundary quality degrades faster than region detection, because the mask has to be right everywhere while the box only has to be roughly right.

Being explicit that you tracked counts rather than mask quality on the CCTV feed is worth more than a claim you could not back, and it points at a cheap next step. You do not need a fully labelled feed to say something defensible. Labelling masks on twenty sampled frames covering the hard cases — dusk, rain, the tightest cluster at the light — gives you a small but honest number, and these failure modes usually show up long before the sample size gets respectable.

That's a genuinely cheap way to get a real signal instead of an assumption, thanks. Sampling the hard cases specifically (dusk, rain, the tightest cluster) rather than a random 20 frames makes sense too — that's exactly where a mask is most likely to fall apart first. I don't have a concrete timeline for it right now, but it's a solid, low-effort follow-up candidate if I revisit this project.

That follow-up can stay small, but I would label it a stress sample rather than a representative accuracy estimate. Selecting only dusk, rain, and tight clusters is useful for finding failures; it cannot tell us how often those failures occur across the full feed.

I should also qualify my earlier interpretation: the 0.604 versus 0.668 aggregate gap alone does not establish that occlusion caused it. Record occlusion and image conditions with each annotated frame, then compare the subsets. That turns the plausible explanation into something you can actually test.