Продвинутый поиск изображений насекомых в Сети

To answer, we must try. And without thinking, hardly much easier.

Also I suspect that not dyuzhe help. It was too much inadequate definitions of the network and look there, "the pearl" ... There are sites that you trust, that they as a rule, you use. But Yuri is right - we must try to speak confidently.

What better time to see one - it is of course ...

And what if to compile a catalog of reliable online sources and make sure that the server was searching for an image only in them? The volume of the material, of course, will be smaller, but the definition will significantly by 99% and foreign is not garbage. Make one more similarity Yandex or Google makes no sense, such as search engines and dofiga and more.By the way, in the same way you can fine-tune the search for any other material - books, articles, films and so on..

Oh, John! No internet in the completely reliable. As always there is any doubt in many cases. You just have to infinitesimal to understand at least in the limited taxonomic group to evaluate the accuracy of the images in the network. And the number of species, surely determined by the pictures, not quite large in relation to their total number.

As I wrote in startposte, item 3 - the problem with other methods of solutions. Here the task to overcome the claims. 1 and 2, i.e. that the system correctly determines what kind / taxa shows According to upload a photo or the author's website . And they are wrong or not is, of course, is not defined.So Ivan thinks as a whole in the right direction (about the restrictions sites for indexing).

Of course, no site is immune from mistakes. I had a case where my photo even Zina incorrectly redefined, although initially I was a correct definition. But my method will reduce the number of bugs and debris to an acceptable minimum.

By the way, it will be possible to do so, he has there and links to photos from other sites if there were to have any kind of lacked a photo from our site.
And, perhaps, if the Internet among the illustrations in mind there are class "super" image, you will need to try to make sure that they are bound to fall into the top selections.Because such a photo as, for example, here: http://www.zin.ru/animalia/coleoptera/rus/cicflams.htm, in my view, should be seen by all in the first place.

Ivan! And among what sources you are going to look for something? Rights Eugene, saying that they simply do not. I talked to entomologists, so they say that all sites are outdated, that has not been updated for a long time (for the butterflies). What very few sites on which you can somehow navigate, but you must know what you're dealing .....

I repeat. We consider the problem in items. 1 and 2. P. 3 do not consider.

Sources photo? Diptera.info, Carabidae of the world, Atlas of Russian beetles ... I can continue listing. They are very much! Besides large sites where tens of thousands of photos, there is still a lot of medium-sized, but also with good material, type and http://coleop123.narod.ru https://www.flickr.com/photos/coleoptera-us, there is also the definition of accuracy is up to 99% inclusive.It would be very makroid in this collection enter, there a lot of photos, but - there is the definition of a mess.

No, of course, and you can add to the list makroid ... But then someone will have to bring it in order for the benefit of our site. To our search engine you can see photos from makroida properly defined.

Lord, here, on the site is the theme for the sources that infinitesimal correct: http: // insecta.pro / ru / community / 2328
It was created more than five years ago and updated periodically :)
There are already many sources infa on which accurate and relevant ( "manual selection") :)
If we intended to fence its own image search - something on the links, you can gather a lot of useful things.offtopic:
Peter, by the way the topic is already time to rename: the site has changed, and in fact is irrelevant: "" kit "(various useful links for butterflies " ..
Who will approach: "" kit "(various useful links for insect " :).

Basil, from the topic pocherpyvat sources, even for the first test version of the search; -)

Lord, to test a preliminary version of the image search: http://ist.insecta.pro
Who is looking at ~ 750,000 photos. Soon we will add the same amount.
Remarks:
1. This is not a gallery. It is the search for the pictures stored on various websites (including on the insects).
2. Now without looking tree taxa and without synonyms.In name only as they are in a photo. Those. OK, you can look on the subspecies, species and genera.
3. When searching for the name of the form must be entered along with the family.
4. The sample pictures in the lower right corner of some images are icons - it's half the individual / individuals and the question mark, if the definition is in doubt.
5.Clicking on Fautua it increases, there is more information about it and a link where you can go to the page where it is located.
6. An increase in ph can browse them using the arrows or prevyuh in the bottom of the screen.
7. Pay attention to the number of results (they beat by 30 pcs. On page in large samples).
8.Who mobile internet - now this thing is pretty zhruschey traffic since in trial prevyuhi actually larger than they appear on the site.
Examples of valid requests in taxa:
Papilio - show all species and subspecies of Papilio, and foty identified to genus / subgenus of Papilio.
Papilio ma - all show that begins at Papilio ma, including swallowtail. Pap - all show that starts on the Pap.
In geography can enter anything you want. Unlike taxa, there is no search mask request *, and mask * Request *. In addition, you can search just all foty any taxa for a given location (enter a query only in the field of geography, but not of a taxon).Example: You can enter in geography "nizhegor" , to try and find you are interested in foty taxon of the Nizhny Novgorod region.

Geography is not specified for all images.
The first page autosubstitution will work, but are not necessary to use (as opposed to other fields on the site). On the page with fotami avtopodstanovok will not.I am waiting for your commentary, that ponra that not ponra.

ZY The project is interesting and specialists for other groups (not only for insects), so gradually add other organisms in the search.

Works, geography, too. The issue there are also some plants, and other members, such as cancers =) Well, again, I remind you that all vyshenapisannoe as a hint should be somewhere on the same page of the search place. For many nuances; himself search activity rather "creative")
I would still like to see a link to the original page.As long as I get there I can not. Visible to photograph.

Yes, crayfish there too, that's fine: -) tips on the search page will do. Link to this page origialnuyu have mandatory under each photo (zoom in and see the description below).
Yes another. Somewhere given geography in Russian, somewhere in the English somewhere in both. Consider the search (especially in foreign locations).

And further. I recall. Tree taxa currently is is ignored. Those. upon request, for example, there are only tribe foty defined to this tribe, but not all child foty taxa of the tribe!

Hmmm ... This is all very interesting ... nothing like I have not seen. Finder at this stage shows good results, although the interface is underfulfilled. In the future, it is worth adding the following things:
- The possibility of sorting the image size, sources, by type, on the accuracy of the determination at step / floor.- Ability to change the definition of photography as part of our server (if it turns out that there is, on the site, incorrect definition). That is, on the site it will remain the same, and in a search engine we have, it will be listed under the true, from our point of view, the definition.
- Ability to comment on the photo.- Date / Time Auto Substitute (according to the data from the website or from the photo properties, and in the priority should be the dating site, because if it is a photo of collection specimen, made in the winter of chamber conditions, naturally, in the properties of the photo will be designated date created file and not the date of the capture of an insect.- Ability to view full-size image, like in Yandex (after the introduction of full-size view directly on our website).

More can be added here is altogether uncertain picture. Because sometimes these stories happen ... here recently described a new kind of dragonfly. We found it by accident while watching a movie. http://www.naturalis.nl/en/news/onderzoek/dragonfly-sir-david-90. And most interesting is that it is - the most natural there the trivial, and the Internet has plenty of her photo hung long before that.But they had never noticed because they are placed anywhere, anyhow, and, of course, no one tried to identify them, so to them it was almost impossible to get to the bottom. If all this chaos to order using our search engine, the search for a new kind of photo will be much easier.

Yes. And another problem: photos clones. If picture author lecturing their creations on the two sites, and more, in a search engine, they are displayed all. I think in the future will need to do both on Yandex - where all clones are automatically grouped and displayed in the search engines in a single copy.Only this will require a special program that will recognize these clones. And, perhaps, a more powerful server.

I found a bug here: https://insecta.pro/is?cap=&geo=Бразилия
There are some images are not displayed when you try to view the site gives the following: "The requested content can not be loaded Please try again later.." In this case the source is working. And, most interestingly, all the photos that can not be loaded, from our website, with photos of other sites all right.

Sort and filter these parameters will, yes.
Commenting / definition will be too.
The date / time the robot collects now, but the data is not as much as geography. But we will also publish.
Now about and so it made as to Yandex. If polnorazmer more - and Yandex, and we refer to the source.Undefined - yes, in the future it is possible, but it is of secondary importance compared to the accurate collection of data that is defined.
About clones - by itself, will be adjusted. But this is when we will be on site at least the image preview (both on the same Yandeksegugle). Now we do not store any sort of graphic data.
With glitches - I know. Repaired.

In Yandex you can not just go to the page with the image, but also to go to the image itself, and our search engine is doing only the first.Add to the future second clearly makes sense - to view the photo will be more convenient (especially if the source photos are sandwiched into small frames with a large size of the property in the Carabidae of the World), or if you want to quickly scan a lot of photos from different sites (for scientists faster and easier -.. the better, that is, to save time).Especially because this feature is available in all top search engines. More precisely, it is all I know of the search engines.

Yes, it's hard to do.

1. Search engine moved to a new address: http://ist.insecta.pro .
2. Now it is possible for the selection, on the floor of the stage, in terms of accuracy (but keep in mind that the values ​​of the corresponding properties are listed is not for all images).
3. The selection of the geography of the second field appeared. Now you can select on the basis of "or."For example, "ekaterinburg" or "Sverdlovsk".
4. The updated interface with pop-up windows and using / reference for each item, as well as for the page as a whole (see. Question marks in the circles).
5. "including unknown" can be assigned to a property in the parameters.For example, you can select photos and pictures of males, for whom sex is not specified, but you can select photos males ONLY.
6. The minimum length of a request for names of taxa, and geography - 3 characters.
7. In the absence of a request for the page displays a random selection of photos.

That is, it is as if we separate from the site will be?

The selection does not always work. Search by name Cicindela campestris species and geography (introduced their area) - it's gone. And when I tried to find them paired pictures - got an empty result, even though I have two pair of mating and there pointed out that this is a pair.

That is it we like to separate from the site will be?
No, it will be combined, just a temporary version of the test would be correct to do is independent.

Regarding the paired photons and all indications floor - yes, there was a glitch with the map now fixed.

I found another bug brutal. On request Phlebodes confixa entire collection just empty, without a photo. I think here it is necessary to improve the program itself, which is the image index.

Yes, the site hosting the foty a portion of the photons are reference curves. Those. the site itself does not seem to know that they are curves. And the robot has not yet been able to determine this. But gradually we teach.
ZY And so glaring problem - because we do not store prevyuhi. But this is temporary.In the standard mode of operation is similar to Yandex / google - prevyuhi store at home, for polnorazmerom helmet on site.

Fixed a bug when foty with Insectoids, who did not have a major version (only 320) is not displayed in the search.

1. At the moment, indexed and searchable ~ 2.400.000 photos.
2. Search is now possible not only for the full title, but also a brief (and will work Papilio machaon, and machaon).
3. Conclusion prevyuh became somewhat less by weight.

Trees, too, will soon connect (only for insects).

I tried to search. Found 28 images, but images of all 10 shows. No matter how hard he tried. Although when it updates the page pop-up images and other, but then go off. And the pictures themselves must be active? When you click on them nothing happens. (Opera browser).

Alexander, reset link on the search results and see what's wrong. We must show all 28 and, by itself, should all be active.

http://ist.insecta.pro/?cap=Aporia+leucodice&geof=&geos=

Very strange. I tried and at the Opera, and the Opera, all the rules opens. It is possible to try the same thing in Opera Turbo mode and in which the thread alternative browser?

Everything works. Of Firefox.

I've even tried by Vivaldi: -)

Alexander, try a different browser, and / or reinstall / upgrade your. Most likely, in the software "dog fumbled" =)

Added a tree (from insects). Now, if you enter the query "Papilionidae", will be displayed not only photos, butterfly which opredelnie to this family, but foty with butterflies all child taxa (of those that exist in the tree). As soon as the tree from insects, then it works only for insects. Trying tree can be on the order level and below.Added a few thousand British and Russian names, which is now also possible to search for.

I found some strange blotches. Request - Hesperiidae, the second page.

And, everything I knew. There, on the site not only shows the photo, but also scans books and maps with locations of finds.

And there is. The robot does not know this phot copy or the original description or something.

In the distant (or not so distant) future, it would be good to teach it to recognize text from pictures.

What for? In most cases, a useful part of the text will be duplicated in an explicit form on the same page.

Well ... I think in the future, when there will be a similar search engine according to the documents, these photos will need to be placed there, it's exclusive material, all scientists of the first description chase at breakneck speed.

Yes anybody for nothing chases. Now holotype is possible through an Internet forward. It previously used email or drove themselves.

Well then fine.

But some texts here and? .. I do not catch up ...