Talk:Standard test image

History[edit]

This article really needs some sources that describe who used these test images. I have never seen Masuda anywhere before, and I couldn't find it on Google.... ~MDD 46 96 00:36, 26 January 2006 (UTC)[reply]

I agree - that would be a good extension of this article. As for the Matsuda image, I know it was named for the researcher who first used it. His name is Ichiro Matsuda and there's a list of his publications at http://itohws03.ee.noda.sut.ac.jp/~matsuda/mrp/ and an email address. I believe that the image was first used by him, then by other researchers. I don't know which of the publications use it, but we could probably find out by emailing him. The Lena image is pretty well documented at Lenna. I am not a specialist in the field, I just used the test images for a project I worked on, and tried to gather as many as possible, then decided to share the knowledge. I found all of them on the web, including the Matsuda images, and looking for them again, you are right, they are gone. I don't know why. PAR 04:49, 26 January 2006 (UTC)[reply]

I noticed that most of the images had been uploaded in 8bit (actually, all of them except Lenna and Peppers)... I had most of them in 24bit png (I presume they're good quality, though I can't remember where I got them from exatly). The only ones I didn't have are Masuda1 and Masuda2 which I do remember seeing a few years ago. The two Matsuda images on the page are currently in 8bit (256 colors). Anyone has them in 24bit? Petit 03:48, 19 February 2006 (UTC)[reply]

It seems to me that the 24 bit pictures are preferable, as long as they are the same size. I think the size needs to be the same for comparisons. If you have 24 bit versions of the same size, could you replace the 8 bit versions? PAR 04:58, 19 February 2006 (UTC)[reply]

I uploaded the images and changed the links. Most of them were the same size. A few were slightly larger (787x576). I think some people stretched them to 720x576 probably because 787 is a weird number: it's not divisible by 16... who knows... Petit 07:11, 20 February 2006 (UTC)[reply]

Image gallery[edit]

This gallery is pretty much a massive blatant violation of both copyright law and wikipedia policy. We've got these images here under a rationale of fair use, which requires that we either be using them to illustrate something that we couldn't illustrate otherwise or else that we're making some kind of critical commentary on the images themselves. Policy-wise, Wikipedia is not a mirror or a repository of links, images, or media files. A couple of typical test images to illustrate what's normally used, and historically important ones like lenna make sense to include, but we shouldn't have a gallery of every copyrighted image that's been appropriated by the imaging field for their own purposes. Night Gyr 07:39, 17 March 2006 (UTC)[reply]

Why don't we just provide a link to [1], which has a much more extensive collection of test images, and is meant to serve that purpose, instead of hosting them all ourselves? Night Gyr 07:49, 17 March 2006 (UTC)[reply]

It does seem to violate the "no repository of images" stipulation, but providing a link to an external repository is not really good either, because external links have a habit of changing or evaporating. Lets not remove these until we have examined the alternatives. Some alternatives:

Add some content for each image, such as a list of journal articles in which it has appeared.
Is there an image-version of Wikisource? If not there should be.

PAR 14:31, 17 March 2006 (UTC)[reply]

There is Commons, but they only want free images, not fair use ones. You still need to cover the copyright issues. Night Gyr 15:07, 17 March 2006 (UTC)[reply]

I have removed the gallery per WP:IUP#Photo_montages. If someone were to find a free image that would be suitable for inclusion. --Oden 21:05, 12 January 2007 (UTC)[reply]

You need to understand that these are standardized test images. They are not replaceable. They are to be used as presented, pixel by pixel, to test image processing algorithms. If you change one pixel in a standardized test image, it is no longer a standardized test image. This is not a gallery of pretty pictures, it is a data set that researchers can use to test image processing algorithms. PAR 23:46, 12 January 2007 (UTC)[reply]

Fair use criterion #3 says:

"The amount of copyrighted work used should be as little as possible. Low-resolution images should be used instead of high-resolution images (especially images that are so high-resolution that they could be used for piracy). Do not use multiple images or media clips if one will serve the purpose adequately."

The gallery/montage consisted of 15 high-resolution fair use images. None of the images were placed adjacent to article text. Fair use criterion #3 says:

"The material must contribute significantly to the article (e.g. identify the subject of an article, or specifically illustrate relevant points or sections within the text) and must not serve a purely decorative purpose. "

None of the images are discussed in great detail the article. I have removed them on these grounds. --Oden 01:21, 13 January 2007 (UTC)[reply]

Also, Wikipedia is not a mirror or a repository of links, images, or media files. Any fair use images will need to be resized to thumbnail size in order to comply with our fair use policy, which will render them useless as standardized test images.

The article on Lenna is an example of a appropriate use of a fair use image. --Oden 01:47, 13 January 2007 (UTC)[reply]

Fair use criterion #3 says:

"The amount of copyrighted work used should be as little as possible."

It does not say the amount of copyrighted work used should be zero. These do not fall into the category of simple copyrighted images. They form a special class of images that are ubiquitous in the image processing literature, and need to be treated as such.

Fair use criterion #3 says:

"The material must contribute significantly to the article (e.g. identify the subject of an article, or specifically illustrate relevant points or sections within the text) and must not serve a purely decorative purpose. "

The images contribute significantly, obviously. They are the very subject of the article.

I have restored them on these grounds, including the anti-spam tag, which is completely inappropriate. The links are links to repositories of other images, and are not "promoting" any web site.

To Oden - What modifications to this article do you think would make inclusion of the images appropriate?

PAR 16:42, 13 January 2007 (UTC)[reply]

A fair use image has to meet all of the fair use criteria. If a free replacement could be created then a fair use image is not allowed (fair use criterion #1).
Multiple images are not allowed if one is sufficient (criterion #3)
Fair use images may never be included as part of a photo montage, as their status as being "fair use" depends on their proper use in the context of an article (as part of criticism or analysis) (image use policy).
A fair use image must identify the subject of the article or illustrate certain parts of the text (criterion #8) However, if a free image could be created then identification is not permitted (criterion #1).

Unfortunately the subject of this article is generic and not specific (an article on test images in general can be compared to an article on oil paintings or on coats of arms). The possibility that a fair use image could be replaced by an image available under a free license cannot be ruled out.

I would also like to point out that articles on similar topics like Color calibration, Color management, Linux color management and CMYK color model are illustrated only using free images. The same is true for the Cornell Box, the Utah teapot and the Stanford Bunny. However, removing any images should not detract from the quality of the article text in any article. --Oden 17:54, 13 January 2007 (UTC)[reply]

I'm still trying to get at my last question - what modifications to this article do you think would make inclusion of the images appropriate? In what sense is the Lenna image a reasonable use of a test image while the others are not? If each test image had its own article, with a description of its image-processing advantages/disadvantages and a history of its use in the literature, would you then think that that would be ok? What about a standard test image article with separate sections for each image, each section giving a description of its advantages/disadvantages and history of use in the literature? How about just the way it is, with a second section that gives a list of relevant data for each image in the first section?

You still don't understand - this is not comparable to oil paintings. I cannot download the Mona Lisa, I can only download a digital reproduction of it. In an important sense, these are not digital reproductions of anything. They are a particular specific sequence of binary digits which happen to yield a picture when displayed in a particular manner. They were purposely chosen to display a picture so that quick, informal judgements could be made on the ability of an image processing algorithm to reproduce those binary digits, utilizing the special ability of the human eye/brain to make such judgements. PAR 06:15, 14 January 2007 (UTC)[reply]

There are some severe misunderstandings going on here. Firstly: You can't compare the images of the Cornell Box, the Utah teapot and the Stanford Bunny to these test images. Those three objects are three dimensional test objects. Their shapes are described as long lists of spatial coordinates - there is no 'image' in the (possibly) copyrighted source data. The making of an image from that data set produces a picture that contains none of the original data - just as taking a photograph of a car does not infringe on the copyrights of the blueprints of that car. Hence, the renderings of those can be free even while the objects themselves might not be. However, these Standard Test Images are just that - two-dimensional images that are exactly what they are. We can no more make a free image of Lenna than we could make our own free version of Gillian Ayres' Image:Ayres, Antony and Cleopatra.jpg.

On the other hand, if we change these test images in any way whatever, they cease to be test images - so a reduced resolution version of Lenna, or one with one row of pixels cropped off - or even one converted to JPEG - is in no way useful as a test image and cannot compete with the original - so fair use is highly applicable in these circumstances.

As for the "Wikipedia is not a mirror or a repository of links, images, or media files", we do not seek to mirror these images because reduced resolution versions of them are utterly useless as test images. For the same reason, we are not seeking to be a repository - a repository of reduced resolution versions of test images is also not a repository. So this policy clearly does not apply. However, when we are trying to discuss why a picture of a baboon (called - incorrectly - a mandrill) is useful as a test image, it's almost impossible to describe the image in words. It's much better to show a picture of the test image (note: Not the test image itself - because by reducing it's resolution, we've destroyed that aspect).

However, I don't agree that we should have a simple gallery of these test images - that would be harder to justify under fair use. But if we discuss each image - what it was intended to test, what notable graphics algorithms were developed using it, where it originally came from, what it's limitations as a test image are, what particular aspects of the image are important to the purposes for which it was selected...then there is no reason why we can't put up a reduced resolution version of that image under fair use - and indeed, failing to do so is akin to discussing a work of art without an image of that art. Compare, for example the English wiki article on Andy Warhol with the German article: [2]. Note how pathetically the latter has a photograph of two soup cans because they can't put up a 'fair use' photo of one of Warhol's famous paintings.

SteveBaker 17:44, 14 January 2007 (UTC)[reply]

I understand what a test image is, and that it cannot be substituted for another image. However, our policy does not allow the use of fair use images if a free image could be created in the future. You would have to rule out the possibility that a there never will be a standard test image available under a free license.
A "repository of reduced resolution versions of test images" is still a repository. Images on Wikipedia should (regardless of the license) either accompany the article text or be deleted.
Also, images should never be referenced in the article text per WP:ASR. Article text should be written with the aim of improving the article, and not merely as a way to allow the use of a certain image.
It is also possible, even likely, that the contents of Wikipedia will be reproduced without images by downstream users. Article text which does not make sense without an accompanying image would in that context be useless.
This is not a place to debate the use of fair use images. At the present time fair use images are permitted on Wikipedia under our policies and guidelines, that is really all that matters. --Oden 01:40, 15 January 2007 (UTC)[reply]

Correct, a free replacement will never be available. The only way a free replacement will be available is if the owner releases it as such. In this sense, these images are no different than a Picasso painting. Cburnett 03:41, 15 January 2007 (UTC)[reply]

Actually a free replacement will become available once the copyright expires. However, my concern is that it might be possible to release test images under a free license. One could even be created (by a university department for instance) and released under a free license. All we need is one, to illustrate the article. --Oden 09:12, 15 January 2007 (UTC)[reply]

So your claim is that no image of anything can ever be placed into Wikipedia under fair use because copyright will eventually expire and we'll have a free-use image?!? That's a pretty extreme viewpoint! If you truly believe that, you'd better head over to the fair use policy page and make that clear...good luck with arguing that!

You go on to suggest that maybe a different set of test images could be created and released under some kind of free license. Yes - I'm sure that's possible - and it may even have happened - but that wouldn't be the STANDARD test images. Go to almost any image processing book - any issue of the prestigious SigGraph conference proceedings - almost any academic paper on image processing anywhere - and you'll see these exact test images being used to measure the performance of some piece of software. They are the standard - any other test images would be useless to the image processing community. If I design a new image compression algorithm, the first thing I do is to reach for the test images, run them through my program and see what comes out. Because it's a standard test image, I can compare my results to those obtained by every other algorithm on the planet - I can search back through the literature of countless image processing conferences, papers and books and see how other people's algorithms compressed these images. They represent a standard against which image processing performance can be measured. Because every researcher in the field is intimately familiar with these images, if I have to show some problem in an algorithm, everyone is instantly aware of what went wrong because it produces some change on the standard test images. What you are asking here is something like "Surely someone could come up with an alternative to the Mone Lisa?"...sure, someone could paint a picture of a woman with a wierd smile sitting in that pose and release it under a free license - but could we use it in discussing the Mona Lisa? No! Of course not - we need to talk about that exact image. Same deal here. SteveBaker 13:20, 15 January 2007 (UTC)[reply]

Removal of gallery of copyrighted images[edit]

I have removed a gallery of copyrighted images which might have constituted a copyright infringement. Fair use criterion #2 says: "The material must not be used in a manner that would likely replace the original market role of the original copyrighted media; our use of copyrighted material should not make it so that one no longer needs to purchase the actual product.". These images are apparently sold by the copyright holder ([3] [4]). The hosting of these images on Wikipedia might replace the market role for the original media.--Oden 13:34, 14 January 2007 (UTC)[reply]

Oden, your interpretation of the referenced pages is incorrect. The text of the linked sipi.usc.edu page clearly states: "For customers who do not wish to access the database images via the Internet, USC-SIPI can supply the database images on CD-R." This is clearly a reproduction charge, not for the images, which they make freely available at this page. Dicklyon 02:39, 15 January 2007 (UTC)[reply]

Copyright does not have to be asserted to apply. In fact the opposite is true, copyright has to be released. If these images are public domain there should be no problem requesting copyright permission and forwarding the permission to Wikipedia. The web page in question could also post a clear and unambiguous message on the copyright status of the images in the database. Of course, the copyright for the image Lenna belongs to the photographer and/or Playboy, so the copyright holder there is someone entirely different. --Oden 03:05, 15 January 2007 (UTC)[reply]

Quite. I was just refuting your statement above re the rationale of "The material must not be used in a manner that would likely replace the original market role of the original copyrighted media; our use of copyrighted material should not make it so that one no longer needs to purchase the actual product," which is clearly NOT applicable here, since the images are all distributed for FREE to anyone who wants them. I'm sure Playboy has given some sort of permission for Lenna, but I am not sure of exact details. Dicklyon 03:31, 15 January 2007 (UTC)[reply]

I'm not sure that Playboy has given any such permission to use the Lena image. I wrote a texture compression paper in 2003 and contacted Playboy to see if I could use the image and was told, most definitely, that I could not without paying for it. In the end I used one of my own photographs [5], which, incidentally, is now being used as test data by other graphics researchers. Simon Fenney (talk) 14:18, 2 January 2008 (UTC)[reply]

That's perfectly possible - but it doesn't matter here (other than that we might want to write about it). We can use the reduced resolution version of the Lenna image under fair use criteria alone. Look down the list of criteria in WP:FU:

No free equivalent. -- We are discussing this exact image - clearly no other will do.
Respect for commercial opportunities. -- Nobody is coming here to look at Lenna as porn - so we're OK here.
(a) Minimal usage. -- We only use it here as a part of a list of standard test images. Check! (b) Minimal extent of use. -- We display it at the lowest reasonable resolution. Check!
Previous publication. -- Yes! This is an exceedingly commonly used image outside of Wikipedia.
Content. Non-free content meets general Wikipedia content requirements and is encyclopedic. - Check!
Media-specific policy...images must meet Wikipedia:Image use policy. - Yep!
One-article minimum. -- Yep. Both here and in Lenna
Significance. -- Yep - it would be impossible to describe the image in words.
Restrictions on location. Non-free content is allowed only in articles - OK.
Image description page. -- Yep.

So we're fully covered here. We meet every single criteria required of us to display this image.

FWIW: If you just write to Playboy about re-using one of their images, you're guaranteed to get a form-letter response from their legal people - I bet they get 100 letters like yours every day from people wanting to use their magazine photos elsewhere, and they probably don't even consider any of them - some lowly clerk sends the form letter no matter what. However, that image has been published tens of thousands of times in image processing literature and nobody has ever gotten in trouble because of it. At this point, there is no way Playboy could or would have sued you for using it. Playboy's official position is made clear in our article about the Lenna image:

"Although Playboy is notorious for cracking down on illegal uses of its images, it has decided to overlook the widespread distribution of this particular centerfold." Says Eileen Kent, VP of new media at Playboy: "We decided we should exploit this, because it is a phenomenon."

"overlook" means that they aren't going to sue people for using it...but if you make it bloody obvious by asking them, then they can no longer overlook it and they have to say "no". Hence, since you were stupid enough to ask and they said "no", you are now on very shaky ground. You may now be the only person in the world who can't use this image! You're going to have to find another career path! However, Wikipedia can use it under fair use - so as far as we're concerned here - it doesn't matter a damn what Playboy say.

SteveBaker (talk) 15:24, 2 January 2008 (UTC)[reply]

This problem can be resolved very easily by replacing the images with reduced resolution versions of the originals. Even (say) a 10% reduction would suffice to ensure that one would still need to purchase the original product since in this case, test images are quite utterly 100% useless as test images once they have been tampered with in any way. Even changing one single pixel in the image would suffice to do that - but dropping the resolution by (say) 10% would make it abundantly clear that we've been responsible about doing that. However, I still don't agree with having a gallery of them stuck at the end of the article. If we aren't discussing a particular image - then we don't have solid fair use justifications for including it. However, if this article were any good, it would be discussing each image in the way that Lenna does for that test image - or Utah teapot discusses it's subject. There is plenty to be said about these pictures because they have each been studied in phenomenal detail - and there are references everywhere in computer graphics literature about them. For example, we might say:

Image:Baboon24.png isn't a picture of a baboon at all - it's a Mandrill - but the person who placed that image into the original test image suite didn't know that. It's chosen because of the highly saturated low frequency yellows, reds and cyans around and beneath the nose - and for the high frequencies present in the fur. This poses challenges for image compression algorithms.

Stick in a SigGraph proceedings reference and that kind of discussion is plenty to justify a low resolution version of the 'baboon' test image.

But yeah - no full resolution versions and no images without discussion of them. SteveBaker 18:19, 14 January 2007 (UTC)[reply]

I would like to clarify the following if anyone is under the misconception that copyright does not have to be asserted to apply. The opposite is true: copyright has to be released. See also WP:C. --Oden 02:55, 15 January 2007 (UTC)[reply]

Requesting copyright permission[edit]

Has anyone considered requesting permission to use one or more standard test images under a free license? See Wikipedia:Requesting copyright permission. --Oden 01:42, 15 January 2007 (UTC)[reply]

Regardless of the TFD & IFD, I would like to see this done or attempted! Cburnett 03:28, 15 January 2007 (UTC)[reply]

Probably won't get you anywhere. Here's what the SIPI doc says:

“

Some of the images in the database were scanned from copyrighted material. The image data was generated from this material with the sole intention of using it for research in image processing and related areas. None of these images should be used for commercial purposes unless you have determined that doing so would not violate any copyright. While SIPI does have information on some of the images, the origin of many is unknown. If you have questions about the origins of a particular image or its copyright status, you can contact us. Be advised, that in most cases our answer to your questions will be that we have no useful information.

”

—The preceding unsigned comment was added by Dicklyon (talk • contribs).

Hmm, well then. I suppose it wouldn't hurt to ask even if the answer is "unknown". Cburnett 06:10, 15 January 2007 (UTC)[reply]

Well, Image:Airplane24.png of a YF-16 might have been public domain if it was take by a employee of the US Air Force (compare with Image:YF-16 and YF-17 in flight.jpg). However, if it is a derivative work then we are back to square one again...

If we are keeping one of the images I would suggest Image:Airplane24.png, because it is most likely to be public domain. --Oden 09:04, 15 January 2007 (UTC)[reply]

Expansion idea[edit]

I would like to see each image given its own section with each image getting the following:

What features make this image standard?
Why this one instead of another similar image?
Where has this image been used before and by whom? (references)

I think this would make the fair use claim stronger and make this article much more valuable. Granted, this effort wouldn't be worth it if they survive IFD. Cburnett 03:33, 15 January 2007 (UTC)[reply]

The deletion process is no excuse for abstaining from improving an article. The argument could even be made that deleting the images makes room for more text as well as free images. The article has roughly the same amount of text since it was first created (first edit). It is a stub with fourteen (14) fair use images. We usually don't allow more than one fair use image at the most in an article this short. --Oden 08:55, 15 January 2007 (UTC)[reply]

You can call edit warring "improving the article" but it is not. Edit warring never improves an article, it only disrupts things. Never. Again: there are no free images. You can't just make an image be ubiquitous like Lenna, nor can you make a new painting and call it a Picasso. Cburnett 17:04, 15 January 2007 (UTC)[reply]

External links[edit]

The links in the external links section need to be incorporated into the text of the article as references, otherwise they will be removed as linkspam per WP:EL. --Oden 03:36, 15 January 2007 (UTC)[reply]

You have a real problem with this article, don't you? Cburnett 03:42, 15 January 2007 (UTC)[reply]

That sounds like a personal attack, but I will assume good faith. --Oden 08:47, 15 January 2007 (UTC)[reply]

It was not an attack. Your actions seem to indicate that you have an issue with this article. Cburnett 17:00, 15 January 2007 (UTC)[reply]

The link for "Compression Database — The superset..." appears broken or perhaps the site no longer exists. I am not familiar with the database. Should the link be removed or fixed? RichardJBarbalace (talk) 19:23, 2 June 2018 (UTC)[reply]

No edit warring[edit]

I would like to remind all users here that edit warring is never the way to resolve content disputes. I have blocked Oden for this very reason. An IFD and TFD are in progress to resolve the issue and discussion on this page seems quite civil.

Again, edit warring is never the way to solve things! Cburnett 04:01, 15 January 2007 (UTC)[reply]

I have brought this block up on Wikipedia:Administrators'_noticeboard#Removing_images_on_sight. --Oden 08:52, 15 January 2007 (UTC)[reply]

Fair use rationale for Image:Airplane24.png[edit]

Image:Airplane24.png is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images uploaded after 4 May, 2006, and lacking such an explanation will be deleted one week after they have been uploaded, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot 02:35, 6 September 2007 (UTC)[reply]

Digital?[edit]

What about non-digital standard test images? They're "standard test images" too, I think.--Exidor (talk) 02:29, 6 January 2008 (UTC)[reply]

What non-digital images do you have in mind? Dicklyon (talk) 04:21, 6 January 2008 (UTC)[reply]

I was thinking of a test image I've seen on film, of a woman with a colorful dress. ISTR it was used in the '60s, probably to compare prints to the original master. But I'll be darned if I can find it again.--Exidor (talk) 18:56, 15 May 2008 (UTC)[reply]

Are you maybe thinking of a TV test card? SteveBaker (talk) 03:15, 16 May 2008 (UTC)[reply]

No, I'm sure it was on film.--Exidor (talk) 17:58, 18 May 2008 (UTC)[reply]

History[edit]

Image gallery[edit]

Removal of gallery of copyrighted images[edit]

Suggested merge of Lenna into this article.[edit]

Requesting copyright permission[edit]

Expansion idea[edit]

External links[edit]

No edit warring[edit]

Fair use rationale for Image:Airplane24.png[edit]

Digital?[edit]