Recently google launched its new feature in image search by image means we can search other images by uploading a image in the google search box. How is this possible?
http://images.google.com
Look at WP:Content-based image retrieval. An example of open-source implementation that you can study internal working of is for example GNU Image Finding Tool.
If you click on the "Learn more" link on the page you are referring to, you'll find this explanation
How it works
Google uses computer vision techniques to match your image to other images in the Google Images index and additional image collections. From those matches, we try to generate an accurate "best guess" text description of your image, as well as find other images that have the same content as your search image. Your search results page can show results for that text description as well as related images.
Actually the answer to this lies in the image processing.....in over a decade image processing and computer vision have done great deal of advancement...
search by image uses pixels ...it compares the pixels and matches with image database it contains....its quite similar to what actual tyext search does but there pixels in place of text...
there are various operators like soble operator,etc which help us focus on the important details of the picture being tested and and we we can search on the basis of the important feature of the image.....
Related
I'm not really too familiar with the programming/coding aspect of Computer Vision. What I can tell from a functional perspective is that it's analyzing an image, then outputting tags based on what it sees. The issue is that the Plugin I use in Wordpress doesn't filter the response of that image analysis. It basically takes my API Key and then echoes the response it receives from Computer Vision to display all of the image tags.
That being said, I have a fairly straight-forward yes or no question. Can Computer Vision be set up to only output specific image tags if they are present in the image? If so, where can I find information on how to do this?
I'm looking at the API reference here: https://westus.dev.cognitive.microsoft.com/docs/services/computer-vision-v3-2/operations/56f91f2e778daf14a499f21b
There does not appear to be any setting allowing one to filter the image tags returned by the service. It also appears that the only format it will return a response in is JSON. So the answer would be no.
I am trying to create an index and skill that will allow me to
Index pdfs, multi and single page, and all other types of files,
Extract the Data and make it searchable,
Search for a term say "Cat" and have sections of text where the term appears to be returned, as well as the page number and document name / downloadable URL of the PDF/ image where it was found, a bounding box, would be nice but not necessary.
I am struggling, I have tried text extraction skill, OCR skill, but I am struggling in that the Search term returns the whole, extracted document (100 pages), as text in the file "content"
It's not making much sense to me, the JFK example is outdated.
I have spent 4 days on this, it cannot be that difficult, the documentation is not that helpful either.
I have tied to "build" and index and skillset using the portal tools, but getting a similar result.
any help would be appreciated.
You might want to try the hOCR custom skill, available on GitHub from the Power Skills repository if you prefer to use the hOCR format for bounding boxes, but [the OCR skill](https://learn.microsoft.com/en-us/azure/search/cognitive-search-skill-ocr#sample-text-and-layouttext-output's output) already offers bounding boxes for content. Note that the Power Skills repo also has updated versions of most of the skills used in the JFK sample, including the image store that can help you make pictures of the pages available in your app.
The key to making it work is in the skillset definition.
The JFK skillset has its OCR skill output layoutText.
There is also a custom image store skill that uploads /document/normalized_images/*/data and keeps the resulting URI as imageStoreUri.
Another custom skill transforms the OCR layout results into the HOCR format.
Then a ShaperSkill is aggregating that information under ocrImageMetadata.
In the case of JFK, that information then gets further aggregated under cryptonyms, because that's the main thing the JFK demo is focusing on, and the image metadata is also an output field mapping for /document/hocrDocument/metadata as metadata, which is also indexed. The important point is that all the relevant information is mapped to the indexed fields. As a consequence, the information therein becomes available from index query results.
According to Microsoft ("Image Analysis" in https://techcommunity.microsoft.com/t5/Microsoft-SharePoint-Blog/Enrich-your-SharePoint-Content-with-Intelligence-and-Automation/ba-p/194174, from May 21, 2018), we should be able to search for text within images.
Is this working for you/anyone? If so, I would like to know what you had to do to get it to work.
I have a SharePoint modern team site with PNG images that contain clearly readable text...but search will not find anything. I have requested re-indexing.
I have had a Microsoft Support request (#10638094) open since June 27 with this question/issue, and no one--even after escalation--has been able to answer it.
Based on the article above, it appears that "MediaService" column(s) should be added to the library to support this; however, I can find no such columns in the environment (using PnP export to review).
Naomi Moneypenny and Kathrine Hammervold highlighted this functionality at Ignite 2017 (https://channel9.msdn.com/Events/Ignite/Microsoft-Ignite-Orlando-2017/BRK2181, about 27:00), but it doesn't seem to be available/working (at least not for me).
August 24: So, after research, digging yet further, I have an escalated support ticket at Microsoft (#10638094, unsolved) and there are conversations at https://techcommunity.microsoft.com/t5/Intelligent-Search-Discovery/Search-for-words-in-your-images-in-Office-365/ba-p/135703, https://techcommunity.microsoft.com/t5/Microsoft-SharePoint-Blog/Enrich-your-SharePoint-Content-with-Intelligence-and-Automation/bc-p/236625, and Does Office 365 image search work? If so, how?. I have yet to hear of this functionality working for anyone. I will keep digging, and I will certainly post if I hear anything. J
After some digging, from official it seems already released at the end of 2017. However there is no any related doc or official guide to this Text in image search function.
The 2 way i can think of perform text in image search.
Perform OCR yourself on the image before uploading the image and embed the text in image metadata.
Use support image type like IIRC and TIF that image are recognized.
In your case, you can upload the image and have another column that contains text and apply metadata to the image in a list/ library column.
OneDrive in another hand also has this function. For example, search for things like "cat" and it * should* pull up most pictures you have of cats. Its more likely using tag as label for the image instead of reading the picture it self.
Also, i believe OneNote has its index recognizable text and handwriting. Maybe this can point you to the right directions.
*Microsoft Azure's computer Vision offer service to recognized text in image. Maybe this can help.
"Is this working for you/anyone?" Yes, I responded to this post elsewhere and see it posted here, as well. Unfortunately, I cannot tell you HOW to get it to work or to verify that it is correctly configured. I can only suggest a test for you to see if it is working for you, as it works for me. I have not tested every way in which it could or should work. I have only discovered it working with PNGs I inserted into Wiki Pages in SharePoint Online. Those PNGs are generated using Snag-It to take Screen Captures and I do not see where Snag-It would be doing any OCR on the image to embed anything, etc. OCR is not even in the Snag-It help file, so I believe the PNG files are just simple PNGs. I insert them into the SharePoint Wiki page, which uploads them to the Site Assets library. And, when I search for a word in the image, the image is returned as a result - not the Wiki page. So, suggest you try a simple test of just inserting a PNG with text in it into a Wiki Page and give the index a bit of time to run to see if it works for you.
It seems like the functionality has matured recently. I have been testing it more thoroughly, and I have documented the results in my blog at http://www.collaboration-foundry.com/SharePointImageAnalysis.
Bottom line: It works for me in OneDrive and SharePoint (modern and classis), but I've only seen it work on the out-of-the-box Document content type--which limits custom solutions somewhat.
It's cool functionality when it works. Looking forward to seeing Microsoft build on this.
John
for my job, I'm looking into an idea in which people would use Google Search by Image and use any celebrity photo they find. Google would return the results and then on our end, a there'd be a database of professionals showing how to get that specific look.
I'm assuming this is extremely unlikely to do, based on that users could use ANY photo.
So, is there a way that I could have about 100 or so celebrity photos that Google Image results could compare to and then choose the one that is closest.
Basically:
Drag drop photo of Britney Spears
Google searches with that image
Google's results compare the top images with our 100, and selects the closest match.
User gets to see video of how to get Britney Spears look.
I'm not a programmer, but looking for some API or Search by Image extension that could make this remotely possible for the programmers here at my job. Does something like that (a search by image api) exist? The best I could find was just the support page, which is hardly of any help: http://support.google.com/images/bin/answer.py?hl=en&p=searchbyimagepage&answer=1325808
You can easily search by an existing image by inserting this into your address bar:
https://www.google.com/searchbyimage?site=search&sa=X&image_url=YOUR_IMAGE_URL
Example:
https://www.google.com/searchbyimage?site=search&sa=X&image_url=http://cdn.sstatic.net/Sites/stackoverflow/company/img/logos/so/so-icon.png
Sorry to say, but the Google image API is deprecated:
Important: The Google Image Search API has been officially deprecated as of May 26, 2011. It will continue to work as per our deprecation policy, but the number of requests you may make per day may be limited.
Quite sure there are some alternatives (http://www.tineye.com/ and http://mrisa.mage.me.uk)
Update (2013): There is now Google Custom Search which allows image searches.
These answers are quite obsolete, but the question comes up in searches. So, the Google Vision API has the "web detection" feature that does a reverse image search. First 1000 requests per month are free, $3.50/1000 afterwards.
I think Google Web Detection could be a solution for you. Google moved it permanently from Image search
You can do it via www.images.google.com but only from a browser (lets you upload your own image and compares it to similar).
I'm working on doing it from code (not from browser).
I had the same problem and came up with two solutions:
There are a number of APIs that give reverse image search results nowadays. The ones I used are https://reverseimageapi.com and TinEye.com.
As the selected answer mentions, you can easily scrape this information but will almost certainly need rotating proxies to prevent being banned by the search engine. There are plenty of proxy rotation services (Zyte, Oxylabs, ScrapingBee, etc.) to make you life easier.
I ended up going with option 1 due to the upkeep of scraping search engines and elements changing / breaking.
I am working on a project to digitize approximately 1 million images for which metadata will be added to facilitate search.
Each image is, for example, a page in a dictionary. But not text. Just a static scanned image. OCR is not an option :(
My objective is to emulate the current search procedure which consists of looking up the alphabetical entries till the correct page is found. In absence of machine readable text, I am looking at tagging each page with Dictionary range tag. For Example (Apple-Canada). So if someone searches for "Banana", it should hit the (Apple-Canada) range Tag.
Is this supported in SharePoint out of the box? If not, is there an addon product which provides this functionality or am I looking at building a customized extension?
Any help will be appreciated :)
Installing the IFilter for TIF files is done with a couple of clicks and gives you free OCR along the way. Very good for scanned pages.
On your question though: No, SharePoint does not have any kind of "range" tags or fields. The only vaguely similar thing to what you are requesting is the Thesaurus of the search. There you could define acronyms and synonyms for words and it would actually search for something else. So you could enter Banana but it would actually search for Apple. Some examples here: How to: Customize the Thesaurus in SharePoint Search and Search Server.
Other than that I can only think of a custom implemented search provider giving you the flexibility you need.