Talk:Machine vision/Archive 1

Old stuff needing a section head
This article has much common with Computer vision, should these be combined?

- It may be wise to accept this discrimination, given that it's so widely accepted although quite arbitrary. Once agreed upon this, it might be useful to state it in each article: we should rememember that articles should be aimed for casual readers, seeking for a definition, not for "us" workers in the field. It might also be useful to organize information so that topics that are common to both subjects be placed in a single position, so to optimize updating and consultation, while placing specific and distinctive topics in the relevant articles. --Zava 19:57, 9 December 2005 (UTC)

- In my opinion, no, the articles should not be combined. There is quite a difference between academic study (computer vision) and industrial applications (machine vision). The fields are closely related but distinct. Professionals in computer vision and machine vision attend separate conferences, publish different trade magazines, and so on. No one in the machine vision field refers to it as "computer vision." -- Rethunk


 * I agree. Its a real faux pas to mix these terms up. Seabhcán 11:18, 12 May 2005 (UTC)


 * Maybe we can say, that machine vision is applying computer vision to industrial applications (like robotics or quality control). It would explain, why we can see redundancy on the Wiki-pages, when it comes to more abstract concepts (f.e. in Processing methods: thresholding, edge detection, ...) wedesoft. 23:44, 5 Oct 2005 (BST).
 * I'd agree with that. (I work in this industry.) There is another aspect: Fashion. Researchers use the term most popular at the time. 'Machine Vision' was popular in the late 80-90 because at that time is was the cutting edge to apply it to real life, but it has become less 'cool' since. Today, Computer Vision is more often used in Universities and Machine Vision in Industry, but in reality the terms are interchangeable. —Preceding unsigned comment added by Seabhcan (talk • contribs)

I changed the first paragraph, as I found the recent edit's defintion of machine vision as applying to the "physical world" vague and misleading. Computer vision scientists as well as machine vision scientists and engineers analyze images captured from the real world. I also made a few quick edits to make some phrases more concise (e.g. "have the ability to" --> "can")

Time permitting, I'll come back and tweak the first paragraph again to make it a bit more elegant.--Rethunk 23:41, 24 August 2005 (UTC)

I' m currently doing thesis about vision based inspection system.. what should I used? computer vision or machine vision? confuse,,,?

Ric —Preceding unsigned comment added by 218.186.8.11 (talk) 06:59, 7 March 2008 (UTC) The use of the term "Machine Vision" is pervasive, and defined by it's pervasive use, as messy as that may be. I think that whoever wrote the definition in the opening paragraph got it about 90% perfect. I made an attempt at helping in the other 10%. North8000 (talk) 12:07, 29 November 2009 (UTC)

No nonsense organization of links (linking vs. advertizing)
Let us admit that self-linking is human. I still feel that a WikipediA should ideally provide independent and non-commercial information. On the other hand, commercial links are indeed a useful resource for users, who do seek a definition but may also want to find suppliers and solutions. So I propose that links are re-organized according to a "no-nonsens" and self regulated policy, such as regrouping them in "Institutional and didactic", "Non profit and public domain resources", "Commercial solution providers" (Or similar categories).

--Zava 17:47, 20 December 2005 (UTC)

Machine/Computer Vision
The section 'machine vision' on the page that discriminate between computer and machine vision is wrong. There is no official distinction. BOTH extend to the field of robot/autonomous vehicle vision. —Preceding unsigned comment added by 130.123.225.69 (talk • contribs)


 * I work in this field and I can tell you there is a distinction. Science is to engineering as Computer Vision is to Machine Vision. You can see this in the journals. Journals which mention Machine Vision are for industry and Computer Vision are for researchers. Seabhcán 09:13, 7 March 2006 (UTC)


 * Seabhcán is right. I work in this field as well. Computer vision and machine vision have remained distinct for some time, though there is significant overlap. The distinction is not a matter of the trendiness of terms, nor is the distinction arbitrary, nor does the "proper" usage depend on whether academic departments in universities decide to use the terms "computer vision" or "machine vision" in their course offerings.   -Rethunk 16:28, 8 April 2007 (UTC)


 * Fascinating. If that is correct I have been working in Machine Vision all this time, not Computer Vision after all.194.80.51.227 10:16, 4 June 2007 (UTC)

I found a ref that says: "...two major reasons for doing this—one is to perform experiments, which will help to provide an understanding of the processes of vision and to throw light on biological vision systems; the other is to permit computers to engage in practical applications such as control of machines. The first of these leads to the discipline known as computer vision, while the second leads to machine vision." Encyclopedia of Optical Engineering By Ronald G. Driggers. It might be good to say so and reference it in one or both articles, because it's news to me. Dicklyon 17:43, 8 April 2007 (UTC)


 * Thanks very much for the reference. This formulation of the distinction between computer vision and machine vision should certainly be quoted in the article. Some day I hope to get back to editing the article again. -Rethunk (talk) 10:08, 1 February 2008 (UTC)

Computer Vision
This page sounds more like somebody's specific job description and does not properly introduce the huge scientific field of machine vision, which is a major modern topic of study in most computer science departments of universities around the world —Preceding unsigned comment added by 130.123.225.69 (talk • contribs)

Seabhcán - try putting 'machine vision' into scholar.google.com —Preceding unsigned comment added by 130.123.225.69 (talk • contribs)


 * I do this all the time - whats your point. Seabhcán 11:01, 27 March 2006 (UTC)


 * Everyone is welcome to read further at scholar.google.com. That said, I'll agree again with Seabhcán; the unsigned comment to which he responded lacks a point. There are textbooks hundreds of pages in length that don't provide a "proper" (complete? comprehensive? broad? deep?) introduction to the field of machine vision. This article was an attempt to describe the field as it actually is

yeah i think they are so cibai. -Rethunk 16:28, 8 April 2007 (UTC)

Machine Vision Terminology
Having been in industrial automation for 33 years, and Machine Vision / Computer Vision for 12, I can tell you that in the former 2 fields it is common for heavily used terms to lack clear definitions, have so many "authorities" that there is no real authority to define such. Such is the nature of the beast, being a confluence of a huge range of fields, technologies and other aspects. So, in these fields, discussion on the existence and definition of terms needs to be guided by predominant usage. "Machine Vision" is a heavily used term, and I think that this article and most comments on it in this talk page do a good job of communicating it's working definition.

And so I think that the gist of the unsigned comment by 130.123.225.69 is incorrect. North8000 (talk) 14:39, 23 December 2009 (UTC)

"No Footnotes" Template Should Eventually be Removed
Knowledge in machine vision comes largely from drawing inferences from reading hundreds of low grade articles, and large amounts of direct experience. The Wikipedia model of in=line reference to authoritative sources does not work very well here.

But there is also a high risk (and track record) of abuse in this article of putting items in the reference section for sales and commercial purposes rather than enhancing the article. These should be taken out and kept out and then the "no footnotes" template removed. North8000 (talk) 18:56, 13 March 2010 (UTC)North8000 (talk) 22:14, 13 March 2010 (UTC)

I reverted deletion of 3/4 of the article
This article is on a very important topic (many billions of dollars per year are spent on machine vision) which is difficult to write about in a WP fashion. There is also a lack of sources which tackle the issues described in my previous talk page post. I was asked by the main publisher in this general area to write a book on this topic but declined. It also suffers from a lack of editors, other than those who want to insert their web sites etc. I made only minor contributions, an updating sentence here and there. However, I can say that the material that is in here (written by others) is accurate, and somewhat on-target. Deleting it all / deleting 3/4 of the article (and it's only reference) as the deletion which I reverted would do a large amount of harm and little good. I have not had the time to do the difficult work needed to further improve and source this article, but I may try to devote a bit more. The "further reading" list includes a few main books in this area and can be used as references for the image processing portions of the current text. North8000 (talk) 12:03, 28 October 2010 (UTC)
 * It might be accurate, but it is unsourced, so there is no eay to verify it. I will individually tag each unsourced section and we can go from there. Please meet WP:BURDEN before reinserting unsourced text. In short, please add sources rather than unsourced OR. Novaseminary (talk) 13:43, 28 October 2010 (UTC)
 * 95% of what I'm talking about here is material that is already in the article, and written by other people which is accurate, not yet sourced, and which is sourcable, the latter being difficult for most people due to the nature of this field as described in my previous talk page section post. North8000 (talk) 13:55, 28 October 2010 (UTC)


 * BTW, I did not mean to re-insert that external link, I did it by accident in that somewhat complex revert. Sincerely, North8000 (talk) 13:58, 28 October 2010 (UTC)
 * Which, if you went through and reinserted material while adding sources; that is, focusing on the actual text that was deleted rather than the volume, you would have noticed. You also removed valid tags without addresnig them Anyway, in reply to your 95% comment, source it, or take it out until it is sourced. A short, well-sourced stub could be far more illuminating than what is here now, and it could possibly encourage others to add to it with good sourcing. We don't ignore WP:V because it is difficult. Novaseminary (talk) 14:01, 28 October 2010 (UTC)
 * If you inserted tags simultaneously with deleting 3/4 of the article, then they might have been accidentally caught up in the revert of the mass deletion. North8000 (talk) 15:41, 28 October 2010 (UTC)
 * I did, and, e.g., per the MOS I unbolded the "MV" abbreviation, etc. I was careful, if bold, in my edit. My criticism of your undo is that it was not. Novaseminary (talk) 15:52, 28 October 2010 (UTC)

Processing methods section
This section needs sources supporting the proposition that MV uses the paricular techniques listed and proper context. While it is good to describe what each technique is and source it, the primary need as I see it is to source the fact that a given process is even relevant to MV in the first place. Of course, the processes should not just be listed in bullet format per the MOS. And, integrating them into prose will also help contextualize the material. Novaseminary (talk) 14:34, 28 October 2010 (UTC)


 * I never edited this article much, and just came back to it and edited and added cites because you just deleted 3/4 of the article, which would have been a big loss for Wikipedia and it's readers. It seems like you are trying to put me in the position as the advocate for / person who has to answer for the article while you pick at it. If you have an interest in this article, why don't you instead genuinely try to work on it? North8000 (talk) 15:38, 28 October 2010 (UTC)
 * I did. I added the only source that actually gives a scope and asserts notability (noting the industry is a $1.5 billion industry). It is easier to grow in a useful, encyclopedic way when the cruft, OR, etc., has been removed. And I only asked you to defend the text you (re)inserted into the article. If you can't or won't, that is no problem,, but then you shouldn't reinsert text. Novaseminary (talk) 15:49, 28 October 2010 (UTC)
 * I just undid your undiscussed unconsensused deletion of 3/4 of the article, wiping out 7 years of work by a range of editors.  Unless you are going to call every word of that  3/4 of the article as something I "inserted", the only thing that I have inserted today or in recent history is citations.  North8000 (talk) 16:33, 28 October 2010 (UTC)
 * I stubbified an OR essay, because I thought the stub more consistent with WP policy and that a stub could serve as the basis for the article to grow in compliance with WP guidelines and policies. You obviously felt the old version was better than a stub. That is what I asked you defend (the whole WP:BRD process). But while we are talking about citations you inserted, with this edit I filled-in the citation you added with this edit, but because of my desire to not violate 3RR, I won't change the template to the correct "cite journal" template (from the current "cite web" template) until 24 hours has run. You could of course do so now, and also explain why you added the source and example and how the source and example are relevant to the section. It seems a bit overly specific to me, possibly violating WP:UNDUE. But then, that whole section is problematic IMO, discussing "typical" processes. Novaseminary (talk) 16:48, 28 October 2010 (UTC)
 * Not sure I follow the second half of your post, but any such changes you would like to make I personally would not consider to be a reversion and would say so if the question were to arise. The format of that section was a list of machine vision processing methods, and neural nets are one of them, not covered by any of the others. This entry is more general/broader than some of the other listings there (such as bar code reading and OCR). North8000 (talk) 17:19, 28 October 2010 (UTC)
 * Fair enough, but why should "bar code reader" and "OCR" be there, either? We could thoroughly (but briefly) describe and completely source either, but that doesn't mean either belongs on this list (or that this list is proper to begin with). Do we have a source that indicates what processes are most often used, or which are representative? If not, this section is WP:SYNTH/WP:OR if we as editors coble together sources only discussing a particular process and come up with the conclusion that is the first sentence of the section (now appearing as "Commercial and open source machine vision software packages typically include a number of different digital image processing techniques such as the following". This sort of list almost invariable is OR unless based on a source that attempts to document the fact. Novaseminary (talk) 17:49, 28 October 2010 (UTC)
 * Due to the nature of this topic juxtaposed with WP policies/guidelines at the granular level, it would be very time consuming (more than I have) to pursue this with you. Other than a sense of duty to prevent your mass deletions of 3/4 of the article and 7 years of work by many editors, I have no attachment to any part of this article. Have at it! Just please be constructive and build vs. just tagging/nitpicking/deleting.   Sincerely, North8000 (talk) 22:53, 28 October 2010 (UTC)

Copied text?
With this edit I removed a paragraph that I thought was a clear copyright violation because it appeared verbatim on page ten of this 2007 dissertation by Hongsheng Song (Machine Vision Recognition of Three-Dimensional Specular Surface for Gas Tungsten Arc Weld Pool, available here, abstract here). But I think I may have been wrong. It appears the paragraph in question was added to this Wikipedia article in July 2005 with this edit by User:Lchatburn. Does anyone have any thoughts on this? The paragraph was unsourced and not very useful anyway, so I would be inclined to leave it out, but the copying seems to have gone the other way. (Also footnote 19 of the abstract cites this Wikipedia article, but does not note that the text is a quote, and other text on page 9 seems copied from this article, too.) Since it is not, in retrospect, a clear copyright violation on the part of the Wikipedia editor, I will self-revert since I am arguably at or near the WP:3RR limit currently. Novaseminary (talk) 01:15, 29 October 2010 (UTC)
 * Aside from a sentence or 2 being a little off-target, I think that it's accurate and useful. I don't have knowledge or opinion about where it came from. North8000 (talk) 01:32, 29 October 2010 (UTC)
 * Well, it would need a source... other than this dissertation! Novaseminary (talk) 01:35, 29 October 2010 (UTC)
 * It also looks like this dissertation has text copied verbatim from Gas tungsten arc welding on page 30 (hidden WP links and all brought into the PDF). Seems odd. Novaseminary (talk) 01:35, 29 October 2010 (UTC)
 * I wasn't able to get the link to work and so I wasn't able to look at it. Maybe someone was using WP to create a paper.   Sincerely, North8000 (talk) 02:43, 29 October 2010 (UTC)

Please discuss on talk page before removing large amounts of material
Novaseminary, you keep gutting the article rather than editing it. Please discuss potential/proposed removals of large amounts of material first in talk. Tag any items that you feel need a cite. North8000 (talk) 13:48, 5 November 2010 (UTC)
 * All of the facts need citations in this section. Most importantly, even in the last version I edited, there needs to be a citation that puts the variosu processes in context. Without that, we have no source supporting whcih processes are critical, which are unique, etc., and this is a context-free bulleted list. Write paragraphs (I have tried to). Novaseminary (talk) 13:50, 5 November 2010 (UTC)
 * "Critical" or "Unique" is not a requirement for inclusion. North8000 (talk) 13:54, 5 November 2010 (UTC)
 * Let's keep the discussion in the section above (below, after refactoring), or feel free to refactor it in this section. Novaseminary (talk) 14:01, 5 November 2010 (UTC)

Proposed plan to expand / improve
Expand the bullets in the "Processing Methods" section then reorganize the material into paragraphs or possibly into subsections. North8000 (talk) 12:56, 5 November 2010 (UTC)
 * Get the sources first, especially sources that explain which processes are used most often, less, often, etc. Then insert new prose, not bulleted lists, per MOS. This is not a technical manual or white paper. Novaseminary (talk) 13:37, 5 November 2010 (UTC)
 * Sourcing definition of "more or less often" is not a requirement for inclusion of cited material. Further, "bulleting" (as that section has been for years) is not grounds for removal of material. North8000 (talk) 13:53, 5 November 2010 (UTC)
 * Please decide under which of the sections you have created you would like to discuss this, and stop adding new sections for the same discussion.. This section is fine by me. I did not delete most of the bulleted material. I moved it into sentences. And context is required to avoid WP:UNDUE concerns. We could write an entire section about Neural Net Processing (with the source you added previously, and one of the few bullets I removed, with this edit), but with other more universal processes merely being listed, the article would be out of whack. I'm not trying to gut the article. (I've added all but one of the sources in it). But it should grow in such a way that the general concepts are described (in prose) first. Perhaps it would be most helpful to work on the Applications section so people know more why they should care about this topic in the first place. Novaseminary (talk) 14:00, 5 November 2010 (UTC)
 * Also, if your company is invloved with any of these particular processes (you've noted above your 12 years in the industry), that would be a reason to step back per WP:COI. Novaseminary (talk) 14:02, 5 November 2010 (UTC)
 * Those are general processes. Knowledge does not equate to COI.  COI is you following me around for psychological reasons and trashing articles wherever I'm there and there is a small amount of (or no) other editors, or stubbornly (for psychological reasons) working back to the gutting of this article 1-2 weeks ago which I reverted.  Tomorrow I plan to restore the version of that section to the version this morning (prior to now) with the fullest content. The extent of the damage that you have done is to great to handle it any other way.   With that in mind, may I suggest storing up any constructive edits for that section until after that and put them in then.      Then, if you again gut this /  remove large amounts of material, I am going to request more eyes on this, probably  via / starting with  a RFC. North8000 (talk) 14:21, 5 November 2010 (UTC)
 * I have already added more sources and context. What you have just threatened to do is the epitomy of edit warring and how you said it it the epitomy of personal attacking; both can get you blocked. Please be constructive. Add contextualized facts and sources. Remember, none of us owns the article. Novaseminary (talk) 14:35, 5 November 2010 (UTC)
 * Several more mis-characterizations. I really wanted to do a RFC on your mass deletions right away, but went to the "restore them once more and then RFC if still not settled" plan in case you wanted to avoid an RFC on this.  Either way is fine with me.  And, as you can see from recent history here, my objection is to mass deletions, not to more specific editing. North8000 (talk) 15:13, 5 November 2010 (UTC)
 * Continuing discussion from my talk page. You have wiped out a massive amount of material from the section.  You wiped out processes, you wiped out all of the descriptions for all of the processes which remain, you wiped out an important step in the beginning,  and the important summary at the end. What I propose is the majority back in, with you picking particular items to stay out, and some way to get the that doesn't take a huge amount of time.  The best way to do that is restore the section, and then you can delete any particular items that you don't want.    I am not amenable to trying to undo mass undiscussed deletions by having to fight back in a huge amount of material one phrase at a time.North8000 (talk) 16:19, 5 November 2010 (UTC)

But per WP:BURDEN, that is what we must do. You have to show why text I have objected to in good faith belongs. There really is not much left that has been cut, anyway, a few of the bullets remain out for particular reasons, etc., and some of the vague descriptions of kept bullets (especially those with their own articles). As I noted on your talk page, I was surgical in what I removed, going so far as to be sure to keep the good sentences you added after you undid my edits. I left much in here. Every particular thing I thought should be removed (based on policy and guidelines, not whim or my personal preferences), I removed, leaving some I think needs better sourcing with a tag. Nothing more. I did not just delete, I collapsed and summarized, too. And since then, I have added facts, sources, etc., to that and other sections. So, other than returning it to a bulleted list (which is its own problem and proposal to discuss here if you think that is a better way to present the information), as I noted on your talk page, I request that you propose the particular sentences or sentence fragments you want to reinsert (per WP:BURDEN) here, with a brief explanation of why and an idea of what source or sources supports the inclusion. We can go from there. For clarity, I would (this time) create a new section entitled "Proposed reinsertions" or the like, but it can continue here, too. We should actually discuss particular facts/sources/summaries without resorting to the tone of the general past discussion--where we don't actually talk about the particular text we disagree on. Cutting “too much” is not a sufficient reason to reinsert unsourced, possibly POV/COI, OR text. But you might have other, legitimate reasons. Let's hear them. Novaseminary (talk) 17:23, 5 November 2010 (UTC)
 * As a show of good faith, I added back the what you called the "important summary at the end" with this edit. I think these two sentences do not actually summarize appropriately, and as new facts would need a source. And they are terribly written on top of that. I welcome editing to show the relevance of them. Now, all that remains removed are two categories of text. The first are the summaries of each of the listed processes. I'm not opposed to that categorically, as I left some of the summaries (editing for subject/verb agreement, etc), but the others, especially where the concept already has its own article, I took out. They were all clumsy and not helpful to a reader unfamiliar with the topic. The second category of still-out-text category are several particular "examples" of processes. Besides getting cruft like, any particular example would need a source indicating its importance to MV. You have not offered anything to support the proposition that those should be mentioned in favor of others that could. In a totally unrelated aspect of this (for an illustrative example), a MV camera might use a special type of glass, but it might not really matter, and noting that some cameras use some particular type of glass would be accurate, could be sourced, but would not belong in the article. Conversely, indicating that cameras themselves are used is useful, of course. We need to establish that listed processes meet the threshold for inclusion. Novaseminary (talk) 17:38, 5 November 2010 (UTC)
 * So, you take a few minutes to do a mass deletion of the sourced descriptions for each of those items, and then I am supposed to spend may hours "proposing" putting them back in, one by one? North8000 (talk) 18:49, 5 November 2010 (UTC)


 * You can't seriously be arguing that my locating and adding several very good sources, several entirely new facts that highlight both what this article is about and why anyone should care, and citations (including tags added well before the deletions) to WP guidelines and policies the removed text/bullets fails only took a few minutes. But since you do not seem to want to actually talk about any of the remaining processes that remain out of the article as of this currently lastest version, I will help. There are 8 (I recently added back “segmentation”). Here they are, pasted verbatim from the last "undo" you did. In the parenthesis next to the item is at least part of my problem with each, though in an effort at brevity, I for now will not go into detail on any in particular.


 * 1. Filters: modify the image to accentuate desirable attributes (such edges or contrast) and reduce undesirable ones (such as noise in the image) (unsourced)
 * 2. Pixel counting: counts the number of light or dark pixel (above or below a particular threshold) or those with gray scale values in between two defined limits. (unsourced)
 * 3. Neural Net Processing: weighted and self-training multi-variable decision making (questionable source and relevance)
 * 4. Activation of discrete outputs to provide simple results such as pass/fail. (unclear and unsourced)
 * 5. Transmittal of more complex results such as measurement and statistical data over buses, interfaces and networks.(unclear and unsourced)
 * 6. Transmittal of object location and orientation results to industrial robots to guide them. (unclear and unsourced)
 * 7. Color analysis tools for differentiation of features or comparison of color quality to a standard. (unclear and unsourced)
 * 8. Positioning/location: Moves the tools to the object or the object to the tools to adapt to varying position and orientation of the object (has source, but unclear relevance, poorly written)


 * My overall objection to the whole software discussion is that it lacks any real context (hence the tag in the sentence intoducing the software processes). That is, we don’t know based on provided sources which of these are major processes, which are hardly ever used, etc. Everything I have read indicates that there is significant variation in processes (which a reader would only know because ‘’I’’ added and sourced that fact).


 * I welcome anyone’s thoughts and constructive participation.
 * Novaseminary (talk) 19:17, 5 November 2010 (UTC)


 * Thank you for that. I really do mean the next sentence in a  a non-sarcastic way.  This defuses the situation enough to where I can calmly leave this as being a hopeless situation. This is due to all of the usual challenges, plus two that are unique here....one is that you are doing your usual thing here, and the second is that this field is such that some of the much needed / most informative content is inherently synthesis distilled from from hundreds of sources, none of which does that distillation/synthesis.  The few people who know enough to write this well (and source it to norm that prevails elsewhere in WP) can go get paid more than a lawyer to teach it instead of being victimized here while doing it for free. North8000 (talk) 20:04, 5 November 2010 (UTC)
 * Ok, but I don't think it is hopeless. There are textbooks written on the subject and some popular press articles. I bet those would be sufficient to create an informative article for a non-expert, yet educated, person (our target) not steeped in MV history while avoiding cruft. Speaking of history, focusing on how MV has developed through the years could go a long way toward putting everything else into context. It is not like MV was developed in a six month spurt back in 2009. As you know, it developed over the last few decades, really. That sort of historical perspective could make an encyclopedia article especially useful even if a popular science or engineering magazine has more fully written about the current state of the industry/technology (though such a source, even if only providing a snapshot, would be great for this article!). Novaseminary (talk) 20:16, 5 November 2010 (UTC)
 * There are textbooks which cover certain aspects of it very well. especially image processing (which is 1/4 the field)   And there are thousands of articles, most with limitations and issues, but many with tidbits of good information.  But I have never been able to find even one good good book on the overall topic, with the "synthesis"/overview type knowledge needed in this field.  And lord knows I've looked, having taught sub-courses on that topic at a couple of universities....also having been asked to write one by one of the major publishers in this area.  Lastly, I didn't say it would be hopeless to write a good article, just in this current environment. North8000 (talk) 20:34, 5 November 2010 (UTC)

Phoenix
I am going to start reconstructing the article to arise from its current ashes. If the above process starts repeating itself, I will immediately go to a process that brings in more eyes to review this situation. I have much more to say that I did not say (and may I suggest the same) in the interest of keeping this on the high road and focusing on giving this important article a path forward. Sincerely, North8000 (talk) 12:19, 16 June 2011 (UTC)
 * Today's tagging of new section is appropriate and I do not consider it to be a repeat of the above process. North8000 (talk) 15:37, 17 June 2011 (UTC)
 * Novaseminary, that new link you put in on image acquisition under Methods is fine. North8000 (talk) 23:49, 18 June 2011 (UTC)
 * Today's aggressive tag bombing of this heavily referenced article (e.g. "Paris is the capital of France" type halves of already-referenced sentences) is IMO starting to repeat the above. Please stop. North8000 (talk) 18:23, 21 June 2011 (UTC)
 * Please stop adding unsourced material about an industry you have said you have worked in for years. Beyond the normal goal of sourcing everything sufficient to make a Good Article, not sourcing and inserting information based on your personal knowledge makes it difficult to be on guard for (no doubt inadvertant) POV. And for other editors (not North, of course) who might want to feature processes their company provides or sells, sources help us root out COI. Anyway, the answer to "tag bombing" is ballistic missiles full of sources, not musket shots on the talk page. Novaseminary (talk) 20:08, 21 June 2011 (UTC)
 * The material which I added is based on sources and heavily sourced. My previous note remains my statement.  I'll resist the urge to say anything further at this point. North8000 (talk) 21:14, 21 June 2011 (UTC)
 * I'd like to help. The article is much shorter than I remember, but after reading it several times I find the changes reasonable. Some of the information is presented more concisely than it used to be. The article may not be comprehensive enough to satisfy MV experts, but the text is accurate enough and--more importantly--not misleading. I haven't looked closely at the edit history, but since North8000 and Novaseminary have contributed the most recently, kudos to both of you for your work.
 * For the short term a few tweaks to the wording would help. The "Imaging" section in particular strikes me as a bit unclear, but that can be edited gradually.
 * If there are concerns about sourcing, I'd be happy to contribute, especially if there's a particular topic lacking proper references (e.g. wider adoption of GigE and USB cameras). - Rethunk (talk) 05:07, 25 September 2011 (UTC)
 * You'll have to really read the talk section to understand. Novaseminary (who I have a lot of history with) followed me around to have fights at articles where they have no interest or knowledge in. That happened here; they gutted & wrecked the article, I got disgusted and left with it in a wrecked state.  Then, months later I came back later and partially rebuilt it and added some additional material and sources. It still needs a lot of work.  Good material is hard to find and add, since this field itself is a synthesis of many disciplines, and so useful information in this field is also synthesis/30,000 foot (but still very technical) views.  So, since we can't do that, one must find sources that do that. Again. it still needs a lot of work, and happy to have you here. Sincerely, North8000 (talk) 11:50, 25 September 2011 (UTC)
 * Thanks. I started this article, and although I acknowledge I don't own it, I do feel responsible for what happens to it even after so many years have passed.
 * For source material, Nello Zuech's book Understanding and Applying Machine Vision should be of help since it includes a concise history of the industry. I have a hard copy of the book, and most of the chapter (pp. 7 - 20) is available as a Google Books preview. - Rethunk (talk) 03:14, 11 October 2011 (UTC)
 * Sometimes "responsible for" sounds like perceived ownership, but I'm sure you didn't mean it that way. I've had many conversations with Nello, but it was always in the context of him gathering and marketing market info.  Thanks for that info, I'll have a look.  But don't hesitate to put material in yourself.   Sincerely, North8000 (talk) 09:21, 11 October 2011 (UTC)

Addition of AIA material
Nice addition Mecheyes. Being a non-profit trade association relieves some of the "commercial" concern, but what would you think of dialing it back a bit, & wikifying a bit, being a bit wp:undue and borderline commercial? Sincerely, North8000 (talk) 18:55, 22 June 2011 (UTC)
 * I removed the material. It was far too promotional/POV and would need third-party secondary sources to establish its proper weight. If the editor did not intend it as a one shot SPAM-like attempt (and we will see, I guess), perhaps s/he could work on the material in their user space first or ask for help here. Novaseminary (talk) 20:46, 22 June 2011 (UTC)

We need better stuff in these areas

 * The 2nd application sentence is dubious in it's context.  The ref had it as a classification method, not as an overview of where the main uses are. As a result it has a rare bird or two in the list. Also the title of the ref is using a particular manufacturer's product.  We should find another source/statement of main uses and substitute that. North8000 (talk) 12:26, 23 June 2011 (UTC)


 * Market size etc. data is changing rapidly in this field. 2006 info is a little old. North8000 (talk) 12:26, 23 June 2011 (UTC)


 * The "non-standardization" sentence under Methods should eventually go.  In the area that the initial source was discussing this is no longer true.  Nor is it true for MV as a whole, but it is true for many other areas of MV.


 * Need to cover:
 * a bit on intra-system communication/control.  Triggers, control of lighting etc. North8000 (talk) 12:32, 23 June 2011 (UTC)
 * common physical connections, physical layers and protocols for I/O North8000 (talk) 12:32, 23 June 2011 (UTC)
 * Applications (expand) A tough one due to the various metrics that this is described by (vision task, process it's attached to, industry that the process is in) sourced pre-dominance info (e.g. avoiding specialized tangents) North8000 (talk) 12:32, 23 June 2011 (UTC)

Merge proposal with Computer vision

 * The lead sentence here needs rewriting.....I've been leaving that for later.
 * While there is some overlap technologically, the common meanings and uses of these terms are very different. North8000 (talk) 16:00, 22 July 2011 (UTC)
 * I changed the intro. Should help with items expressed in both articles. North8000 (talk) 19:39, 22 July 2011 (UTC)
 * Discussion was centralized at the other article and closed out by others as "no merge". Although I'm with the "they're different" camp, I wanted it to run plenty long to thoroughly examine and settle the question. North8000 (talk) 09:24, 11 October 2011 (UTC)

Removal of direct interface section
Novasemianary, with, in your usual obsessive battling behavior here, you removed the sentence about direct interface cameras, even after I said that it was crucial and that I would get it sourced. This also rendered the previous sentence now false, which was dependent on that sentence. So, you go around wrecking things and the adults have to fix them. I'll put a more limited sourced sentence in there to repair the damage. With this behavior you are getting very close to being the first person I have ever reported for anything. Please stop. North8000 (talk) 20:24, 28 November 2011 (UTC)
 * I take it this edit is the one to which North objects. This sentence has had two tags since June. If it is so critical, I would expect several sources to note as much. Your inability to add even one after months leads me to believe the sentence is OR and only supported by your experience in the industry, or worse, POV. Add an RS that says as much and there is no problem. Your promise to add a source in the future does not meet WP:BURDEN. For what would you report me? My edit is upported by policy. You reinserted unsourced material that had been long-tagged and subsequently removed. Instead of finding and adding a source, you waste our time complaining and threatening here. Novaseminary (talk) 21:03, 28 November 2011 (UTC)
 * I am talking about your overall pattern of obsessive battling behavior (using wiki-lawyering to do it and disguise it), with this article being an emblematic example of it. North8000 (talk) 22:18, 28 November 2011 (UTC)
 * In that case, please stop using this talk page in ways counter to WP:TPG and, here or anywhere, please refrain from making personal attacks (WP:NPA). Novaseminary (talk) 22:48, 28 November 2011 (UTC)
 * The above is an attempt to resolve the dispute here. And discussing behavior is not a personal attack. North8000 (talk) 22:50, 28 November 2011 (UTC)

Well, in that case, would it be ”obssesive battling behavior” if I fix the Dinev cite to italicise the journal and put the article title in quotation marks? Or do I have an irrational phobia or aversion to improperly cited refs? Novaseminary (talk) 23:01, 28 November 2011 (UTC)
 * You always ask about the benign edits but not the battling ones. The latest battling one was adding "than what" to "newer", where the answer is obvious in context. Please stop. North8000 (talk) 23:12, 28 November 2011 (UTC)
 * I don't think it is obvious when (other than reletively) this phenomonon began. Doesn't the source date it? Please stop what? Happy editing! Novaseminary (talk) 23:17, 28 November 2011 (UTC)


 * The meaning is very clear, both by the sentence which it follows and by the insertion in the sentence itself. Newer than the framegrabber method. I am taking that battling tag off.  Please get a consensus here if you want to put it on.   And please stop the aggressive battling type editing.   North8000 (talk) 11:49, 29 November 2011 (UTC)
 * The meaning might be clear to you as somebody who has worked in the industry for more than a dozen years. But it needs to be clear to a non-specialist educated reader. With this edit I have added a time frame. This article should not be a mere catalog of MV techniques, but should walk the ready through the history (with dates!) and, without unnecessary jargon, and tell the reader where the industry (or technology or whatever) is today. Please stop interpreting my demands that this article be well sourced and accessible as ”battling”. It is not me, but WP that demands this. This is all the more so in an article like this that could easily become a POV piece favoring certain techniques of a particular company, or what have you, without a non-specialist even realizing it. This is an encyclopedia, not an industry journal. Novaseminary (talk) 14:49, 29 November 2011 (UTC)
 * The reality (integrated from hundreds of sources) is that prior to approx 2001, all separate-camera machine vision was done using frame-grabbers. Beginning in approx 2001 those direct interface methods (that did not require or use a framegrabber) started becoming available. Over the ensuing decade, those methods have steadily grown to where they constitute the majority of separate-camera vision systems. But that is a summary / vetting  from a combination of hundreds of sources.  Not all in a tidy single Novaseminary-assault-proof single sourcing situation that I am aware of. North8000 (talk) 16:49, 29 November 2011 (UTC)

That would be great to add to the article, just as soon as it can be sourced without WP:SYNTH. You should write an article or primer on machine vision! Novaseminary (talk) 17:06, 29 November 2011 (UTC)
 * Actually the lead editors of the most prominent vision vision magazine twice asked me to write a book which I have declined. (There is a lack of quality "forest for the trees" material in this field, which is part of what makes it hard to put useful material here sourced in a way bulletproof enough to withstand an assault by you.)   Then I must have had a a brain injury and decided to edit here from sources for free and be abused by you.  :-)  North8000 (talk) 17:44, 29 November 2011 (UTC)
 * I'll try to find more in sources to put something better in that sentence. North8000 (talk) 22:09, 29 November 2011 (UTC)
 * In the meantime I think most would agree that it is better to be non-specific than to be misleading, which the "2008" is (about 7 years off). Further, the "more recently" term is the most specific that can be summarized from that particular source without adding unsourced specificity.  Such wording is a common practice in writing in general and in Wikipedia to follow what is and isn't know/sourced respectively. So, anyone wishing to try to re-add tagging to "more recently" should obtain a consensus in talk first. North8000 (talk) 11:49, 1 December 2011 (UTC)
 * Thanks for the primer on writing in general. This field could use help in that area (take the Turek article you cited and I added the URL for, for example-less than elegant writing, in my opinion). Anyway, since we cannot source when particular methods developed, and sources do indicate the methods are available now, let's just focus it in the present. Novaseminary (talk) 15:49, 1 December 2011 (UTC)
 * Now we've lost the sourced and useful fact that the transition occurred due to your battling behavior on that sentence. North8000 (talk) 21:55, 3 December 2011 (UTC)

In case anybody cares, this is the "battling" edit North8000 is talking about. "Behavior" aside, I think the edit improved the article. Novaseminary (talk) 05:33, 4 December 2011 (UTC)
 * The last edit of the edits in that diff and multiple previous edits on that same sentence are the battling part. I'll probably wait until I get more from sources before I fix it. North8000 (talk) 11:21, 4 December 2011 (UTC)

3D material which I added
The 3D material which I added was aggressively messed with by a person with a long history of following me and doing so. I'm going to let the dust settle a bit before undoing the damage. Regarding location, the 3D material is equipment technology and a technique, not an application. Important & informative sourced material on the methods and technologies was also deleted. North8000 (talk) 02:42, 13 May 2012 (UTC)
 * It was aggressively brought into line with the (belatedly) provided source. And the author of the source called it an "application (per the quote), but I've moved it into the methods section. The other material was unsourced and the determination it is important seems to have been made by North, not an RS. Novaseminary (talk) 02:59, 13 May 2012 (UTC)
 * That's erroneous on several levels. The techniques were sourced.  And yes, it is editor discretion to decide that people will want to and expect to find out a bit about how 3D vision works when they come here.   That is not improper as you imply.  Deleting material hours after it is added for being allegedly unsourced in not the norm, it is your usual obsessive aggressive targeted tampering with edits of editors that you have had disagreement with. 3D can also refer to the applications that it is used on, but the material that you tampered with/ messed up was not applications, it is technologies  for accomplishing them. North8000 (talk) 03:09, 13 May 2012 (UTC)


 * In this round of the above-described tampering, you tagged a ref for page numbers. In a previous round of such tampering/ damage, you were the one who removed the page numbers on that reference.  A further example of the above. Stop!  North8000 (talk) 03:16, 13 May 2012 (UTC)
 * Maybe editing is tampering to you; I disagree, of course. Anyway, if I removed page numbers from a citation to a book, I certainly didn't mean to. Novaseminary (talk) 03:20, 13 May 2012 (UTC)
 * No, as you know, your first sentence is not what I meant.  I meant pursuing your obsessive battling behavior (per your long history of doing so) by aggressively tampering with edits of whoever your target is. Usually it is a mix of some good housekeeping edits combined with destruction of the material, and such is the case here.  The final note was illustrative that your battling is just that rather than of merit.  During one round of battling via aggressive editing you removed the page numbers from the reference, and, then, forgetting that you did that, in a later round of aggressive editing you tagged that same reference for lack of page numbers. North8000 (talk) 12:10, 13 May 2012 (UTC)
 * Putting aside North's inaccurate and hurtful characterization of my edits (edits made in an honest effort to insure WP is sourced and not promotional, to maximize the value of the encyclopedia), I for the life of me cannot find where I removed page numbers on this article. Again, if I did, I did not mean to, and I apologize. (Perhaps North was thinking of this edit where I consolidated a reference that had been repeated several times, to be one ref pointing to several pages; but all of the page numbers are still there, just in one ref since they all support the same sentence.) I did find this edit where North removed a page number from a reference that should have a page or pages listed. I am glad we can agree now that a page number, numbers, or range of numbers belongs there. Novaseminary (talk) 19:20, 13 May 2012 (UTC)
 * I wish the first line of your post was the reality; if it was, things would be much better. Removing pages for cites is not optimal but not an apology level issue. The battling behavior illustrated by the reversal is.  I plan to reconstruct the wrecked 3D section.  I'll probably wait 2 weeks as I'll be overseas where web access is a question mark.  Also to separate the genuine improvements from the damage. North8000 (talk) 23:20, 13 May 2012 (UTC)

It would be superb if you would add quality, sourced material placed into proper context and given appropriate weight. If you add another several paragraphs based on personal knowledge, less so. Because, despite your assertion above that WP's inline citation policy "does not work very well here", it remains the policy. It is inappropriate to insert material based on "drawing inferences from reading hundreds of low grade articles, and large amounts of direct experience" as it seems you did yet again with this section. Of course, you know this. It must have been a momentary lapse when you did it, and then another when you called my reworking of it "tampering" that "wrecked" it. Novaseminary (talk) 00:50, 14 May 2012 (UTC)
 * Summerizaiton of sources is what we do here.
 * The source that you added supports the material that you deleted when you wrecked it, not the material that is there. What a mess. I'll fix it when I get back from overseas.  I guess that one plus thing is that you have sprinkled some improvements in there along with the damage.   So the best approach is to sort them out which will take more time than just reverting the damage. If you didn't do so much obsessive battling disguised as editing, you could do a lot of good. North8000 (talk) 11:42, 14 May 2012 (UTC)
 * Yes, but we use inline citations and not one editor's synthesis of "hundreds of low grade articles" and never "direct experience" unless covered in an RS. Maybe you should write an article in an RS... Novaseminary (talk) 18:28, 14 May 2012 (UTC)
 * Much better to edit here for free from sources and get harassed. :-)  North8000 (talk) 18:56, 14 May 2012 (UTC)
 * Even better: cite those sources and don't go beyond those sources so you don't get harassed. Novaseminary (talk) 19:41, 14 May 2012 (UTC)

weasely repetitive lead
The lead gives two very similar definitions and also has an incredibly weasely bit in the middle that says the "definition is difficult to distil". I fixed it but it was reverted Bhny (talk) 03:21, 13 September 2012 (UTC)
 * If the RSs indicate that the definition is not widely settled, the article should reflect that. I do not see how what is there, or what you are saying, fits the description in WP:WEASEL. In fact, the sentence seems sourced pretty well. Novaseminary (talk) 03:25, 13 September 2012 (UTC)

Every article could say "this topic is difficult to define" it's the ultimate in weaselyness. Just because you have sourced something like that doesn't mean it's good. The job of the lead is to define the topic, not grumble about how difficult it is to define. Also I would expect the 2 definitions to be different if there was this difficulty Bhny (talk) 03:36, 13 September 2012 (UTC)
 * I doubt many articles about people, or places, or particular things would face any real definitional problem. More abstract ideas and processes, like machine vision, maybe. But the "weasel" language is straight from a RS (or at least one you have not challenged or contradicted) not my or any other editors OR or interpretation (that would be a problem and violate WP:WEASEL). If a topic is sufficiently hard to define, or not yet defined with consensus among those in the know, why isn't that fact worthy of inclusion if sourced to a RS that found it worthwhile to discuss early in the book (page 5)? Novaseminary (talk) 04:11, 14 September 2012 (UTC)


 * I agree with Novaseminary, which doesn't happen often. :-) Not that the lead couldn't use work, but this aspect is more salient in this field than it is in a more typical one. North8000 (talk) 10:25, 14 September 2012 (UTC)


 * Then expand on these competing definitions. At the moment the lead is- MV is def1 . definitions are difficult but MV is def1 again. Bhny (talk) 16:37, 14 September 2012 (UTC)
 * Agree that we should do that. North8000 (talk) 16:58, 14 September 2012 (UTC)

OK well I made a start by combining the 2 similar definitions. Now we can add these contradicting definitions Bhny (talk) 22:19, 14 September 2012 (UTC)

"Multi-Camera" insertion
I know that the discussion has centered around sourcing, but that insertion is also really not appropriate for that location. There's nothing indicating that it isn't 2D visible light imaging, which is what that enumeration is. Now multi-camera stereographic imaging would be a different story. Sincerely, North8000 (talk) 16:58, 4 March 2013 (UTC)

Unsourced list
Moving this unsourced list here per WP:PRESERVE. Per WP:BURDEN this can be restored when reliable sources are provided:


 * Automated Train Examiner (ATEx) Systems
 * Automatic PCB inspection
 * Wood quality inspection
 * Final inspection of sub-assemblies
 * Engine part inspection
 * Label inspection on products
 * Checking medical devices for defects
 * Final inspection cells
 * Robot guidance and checking orientation of components
 * Packaging Inspection
 * Medical vial inspection
 * Food pack checks
 * Verifying engineered components
 * Wafer Dicing
 * Reading of Serial Numbers
 * Inspection of Saw Blades
 * Inspection of Ball Grid Arrays (BGAs)
 * Surface Inspection
 * Measuring of Spark Plugs
 * Molding Flash Detection
 * Inspection of Punched Sheets
 * 3D Plane Reconstruction with Stereo
 * Pose Verification of Resistors
 * Classification of Non-Woven Fabrics

-- Jytdog (talk) 01:38, 9 February 2017 (UTC)


 * I think that the entries were too specialized anyway. A shorter list of broader areas would be better.  North8000  (talk) 13:38, 9 February 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Machine vision. 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/20120605095256/http://www.designnews.com/document.asp?doc_id=237971 to http://www.designnews.com/document.asp?doc_id=237971

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 12:56, 11 January 2018 (UTC)


 * Looks good although the directions appear faulty. I don't see any "this template". North8000 (talk) 13:43, 11 January 2018 (UTC)