Talk:Straight skeleton

This algorithm can be used to obtain the skeleton of the character in OCR. —Preceding unsigned comment added by Sergey.Radkevich (talk • contribs)

I wouldn't be surprised. If you can find a published reference to that then we can add it to the page. —David Eppstein 19:17, 19 February 2007 (UTC)[reply]

Any reference on Algorithms is welcome RzR 08:09, 19 September 2007 (UTC)[reply]

The 3rd image on the right of the page is absolutely *awful*. Why introduce a confusing shadow to the mix? Why orient the thing in some other direction compared to the other two images? Is the 3D version even the *same* as the other two images? It isn't even certain. Epic fail in terms of the goal of assisting in understanding. It raises more questions than it answers... 24.162.243.254 (talk) 17:34, 5 December 2011 (UTC)[reply]

The straight skeleton was wrong, too. I replaced it with a hopefully better version. Shuber2 (talk) 09:23, 18 September 2012 (UTC)[reply]

Reference 11 "Computing the straight skeletons of a monotone polygon in O(n log n) time" by Das et al. has been shown to be incorrect. Propose to refer to "Straight Skeletons of Monotone Polygons" by Held, Huber et al. instead, in which same time bound is reached.HexTree (talk) 21:20, 15 March 2014 (UTC)[reply]

Ok, I added the new reference. From what it says, it seems that Das et al. are not incorrect, just overly restrictive (they require the input to be in general position, which is not a good assumption for straight skeletons). Or is that an inaccurate description? —David Eppstein (talk) 23:26, 15 March 2014 (UTC)[reply]

Lemma 5, 6, and 7 and their proofs are wrong. Lemmas 5 and 6 can be "repaired" by transforming them into assumptions. Lemma 7 attempts to prove the same wrong statement as Felkel and Obdržálek did. There are trivial examples for which d(\xi, \ell(f)) in this Lemma can be made 0 while d(\xi, \ell(e)) is arbitrary. Shuber2 (talk) 08:40, 16 March 2014 (UTC)[reply]

Algorithm section variables

The variable r is never explained. Is it the number inward corners? — Preceding unsigned comment added by 87.211.13.55 (talk) 09:42, 11 November 2018 (UTC)[reply]

Yes. It was explained, but lower down in the bullet for Cheng and Vigneron. I added another earlier explanation. —David Eppstein (talk) 19:15, 11 November 2018 (UTC)[reply]

I am sorry I did not recognize your changes: I added a whole paragraph to explain the input. Now it is maybe too much? Stefan Huber (talk) 20:13, 11 November 2018 (UTC)[reply]

Probably, but I think having the notation at the top instead of explained in conjunction with the individual time bounds is better. So maybe the later explanations should be removed or merged into the one at the top? —David Eppstein (talk) 20:18, 11 November 2018 (UTC)[reply]

I gave it a shot. Stefan Huber (talk) 20:29, 11 November 2018 (UTC)[reply]