Talk:Active contour model

Untitled
I don't understand why an article with the same name has been deleted. Unless you hold no less than a PhD in Computer Vision, Image Processing, Computer Science or Mathematics, please do *not* delete it.

Snakes, or Active Contours, are one of the most used object object identification methods. For non-scientists, snakes are namely used in the "magic wand" Photoshop(TM) tool, in a very basic version.

Suggest or apply as many modifications as you like: this page is only in a preliminary version, and deserves a long and detailed description.

I am doing a PhD in computer vision and like to enhance this page with out deleting any existing text. I will add more detailed description under different subsections. Please comment. —Preceding unsigned comment added by Nautiyaa (talk • contribs) 12:55, 19 February 2010 (UTC)

Do we really need this gruesome image to illustrate how active contours work? As a bit of an arachnophobe, personally I find it distracting. — Preceding unsigned comment added by 69.91.151.133 (talk) 16:47, 12 January 2012 (UTC)

-> Agree, please find another picture!!! 188.103.125.180 (talk) 21:42, 14 November 2012 (UTC)

I'd love it if someone could derive the termination functional (I can't myself), or at least add references to the relevant vector calculus. 173.247.198.204 (talk) 02:54, 24 February 2013 (UTC)

Not all active contours are snakes!
Please note that not all active contours use a snake model. (And yes, I hold a Computer Science PhD which involved working with non-snake active contours.) Please see the book "Active Contours" available at http://www.robots.ox.ac.uk/~contours/ for reference. 188.103.125.180 (talk) 21:42, 14 November 2012 (UTC)

ERROR: DOES NOT COMPUTE
There is no way I could interprete the formal part without already knowing what it is supposed to mean. The following definitions are missing: Function v(s), v_s(s), v_ss(s), vbar

There is only the definition of the SET of points — Preceding unsigned comment added by 5.158.164.158 (talk) 08:01, 29 April 2013 (UTC)

I believe that the derivation of the internal force is incorrect. It follows from Euler-Lagrange's equation and not by taking the gradient as it is described now. — Preceding unsigned comment added by Herman5387 (talk • contribs) 14:10, 19 August 2013 (UTC)

Readability...
To use an awfully clunky word. I'm new to active contours and was reading this page to give myself a general background before reading primary literature.

There are several problems relating to basic English that seem to need sorting. Examples include clunky sentences that don't make sense until you read them a dozen times. Missing pronouns. Incorrectly used words, e.g. the last sentence of GVF Active Countours, the use of the word especially doesn't actually make sense. I think, if somebody has the time to, reviewing just the English grammar and syntax used would make this article easier and more clear to a novice reader such as myself.

I would offer to make the editorial changes myself but, not being well read on the area, I would feel uncomfortable editing a topic I don't fully understand. — Preceding unsigned comment added by 93.96.60.125 (talk) 10:18, 16 July 2013 (UTC)

Advantages and drawbacks
The article states "Their accuracy is governed by the convergence criteria used in the energy minimization technique" as a drawback.

For me, this clearly fits better to the advantage section, because the tradeoff between accuracy and calculation time can easily be adapted to the needs of the application. — Preceding unsigned comment added by 93.206.202.135 (talk) 03:57, 15 September 2013 (UTC)

Plaigiarizes Article
The section on Termination Functional is lifted word for word from the article "Everything You Wanted to Know About Snakes But Were Afraid to Ask" by Jim Ivins and John Porrill. See page 10 of this PDF. While the article is freely available it states that: Verbatim copying and distribution of this entire document is permitted in any medium, provided this notice is preserved, but changing it is not allowed. — Preceding unsigned comment added by 67.175.245.194 (talk) 10:56, 29 December 2013 (UTC)

The mentioned content is similar to the original paper.

Pseudo code is not really pseudo code
I am not sure if this is a general problem of embedding pseudo-code in Wikipedia, but the 3 bullets are, by my standards, not acceptable as pseudo-code. Pseudo code should enable the reader to easily implement the algorithm. The presented bullets do not do so. There is talk of iterations, then why is there no notion while/for/repeat? This is important as it would clarify on what stopping criteria to use and which steps are in the loop and which are before it.

Moreover, the input/output specification does not exist, one could argue that the Wikipedia article specifies them, but I would still like to note this.

I do need to admit that the Implementation section is quite clear about the gradient descent being used, and added clarity to the original Terzopoulos paper, regarding the update rule. However, the Implementation section suggests that one does an *online*-update of one snaxel (i.e. a point on the snake) and then move to the next snaxel.

Bottom line is that pseudo code should be close to what code looks like in real life, with shortcuts where it would become tedious for the reader, not tedious for the writer. I would suggest some code like this:

input: initial snake (set of snaxels V=v1...vn) output: converged snake (also set of snaxels V=v1...vn) repeat: for v in v1...vn          // apply gradient descent according to $$\nabla E_{snake} (v_i) $$ as defined in the Implementation section v:= $$v-\lambda \nabla E_{ext} (v) $$ end until snake converged w.r.t. some maximum allowed amount of change in the vi's

I am not an expert on the field of active contours, so please verify and correct the code before using it. I feel that the calculation of $$\lambda \nabla E_{ext} (v)$$ should be done between the repeat and the for statement, to have more or less the effect of doing offline updates. But I am not sure about this.

I hope this is useful :)

External links modified
Hello fellow Wikipedians,

I have just modified 2 one external links on Active contour model. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20110716113957/http://www.isbe.man.ac.uk/courses/Computer_Vision/downloads/L11_Snakes.pdf to http://www.isbe.man.ac.uk/courses/Computer_Vision/downloads/L11_Snakes.pdf
 * Added archive https://web.archive.org/web/20101008005512/http://www.isbe.man.ac.uk:80/~bim/software/qsnake_demo/qsnake_demo.html to http://www.isbe.man.ac.uk/~bim/software/qsnake_demo/qsnake_demo.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 17:37, 3 October 2016 (UTC)