pKozaneba2025-08-29
range selection
https://gyazo.com/437c726a381131c4f23ca322691ceb0c
Hmmm, a rectangular selection of connected components would cause overlap, that's true...
Another approach is to divide every 10 squares.
That would eliminate the crossing problem, and it wouldn't look messy.
It happens that "natural clusters" get split up.
In the use case I'm using Kozaneba, a human being is doing the collecting of similar things in the first place.
"I want to put this in one group, this is not."
Because the gap is driven by the decision to
These overlap problems don't happen.
I knew the route and what to do with more than 1,000 cases,
Kozaneba is structured while creating less than 1,000 stickies.
I think it's better not to equate them indiscriminately.
Merging and squaring of overlapping objects
https://gyazo.com/1992204dfafe40f76cd8c0adab2f84f2
Only merging if there is more than 30% overlap now.
This is what happens when you reduce this to 20%.
https://gyazo.com/403217eebe831a46dd705e26132b4691
Maybe it's a chain of connections around here.
https://gyazo.com/23ef63b297d7e1ef990c26a536dd8f8a
---
This page is auto-translated from /nishio/pKozaneba2025-08-29 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.