i think this goes beyond the constant id aspect of it or are you arguing that nearest neighbor is not the only/wrong approach to the problem?
I mean that the nearest neighbour of centroïds is not the only approach to getting constant ids. There are many possible ways of getting constant ids. I just haven't tried any yet.
I completely agree with this, yet I think that there is some confusion.
one thing is the result of the tracking which can give you centroids and anything else. Another is getting those tracking ids and have them be continuous.
What I am getting is that you are obtaining regions of pixels based on some criterion, for each frame
Yes, this is what I already do when I'm doing a multiblob detect without constant ids.
Exactly
and then comparing how they overlap with the previous frame?
That's what I propose but haven't tried.
And this is a separate process that does not necessarily follow from using centroids or anything else.
are you assigning constant id to the objects with maximum overlap?
I haven't done it, but this sounds like what I would do. There might be more creative things to do with those labelled pixels, but I think that the maximum-overlap thing would be something to be tried.
Actually, one can already build an overlap matrix quickly, using [#join] and [#draw_polygon + 1, draw point] (the polygon thing is a misnomer in point-mode.)
well, the problem with maximum overlap is that an object might have moved faster than one would expect and it can show minimum overlap or no overlap and still be a desired match.
This case would make the sampling problem worse. While we might want things that have no overlap to still be matched, this would create new ids constantly...
J