I know I can use google to do an image search by putting in a key word and just scanning images.
Can someone take an image (JPEG or whatever) from an on-line source and try to match that image exactly?
For example if I post up an image of myself -- can someone take that image and search for it in other places on the internet?
Is this highly improbable or would someone need really really specialized soft ware to do this. I don't think that even the police can do this because otherwise they would be able to catch child pornographers or people who trade it more easily.
AFAIK there is no software performing an image based search in the web as you described.
This would demand for fairly time consuming image recognition, pattern matching and/or registration algorithms. Multiple image formats, resolutions (x/y) and color depths must be supported.
Other searches that permit searching for a keyword need a data base or enhanced image tags that describe what can be seen on a particular image, or stuff like painting style (oil, foto, pencil, coal and so on), epoch, artist and so on.
Efforts have been made to allow for a rudimentary content referred search so that a user can specify an image like 'mostly blue' or 'part red, part yellow'.
This report looks quite serious, although it is over one year old. It says pretty much the same as I did, but with more words and mainly with more examples.
The fourth chapter may be of particular interest:
4. Content-based Search Engines
All the image search engines considered so far in this review have been based on text and context strategies or on associated catalogue entries. There have also been a number of attempts to build content-based search engines. Content-Based Image Retrieval (CBIR) considers the characteristics of the image itself, for example its shapes and colours. To date, these attempts have been experimental and generally limited to individual collections.
Well-publicised efforts have included Columbia University's WebSEEk project (
http://persia.ee.columbia.edu:8008/) and IBM's QBIC (Query by Image Content -
http://wwwqbic.almaden.ibm.com/), which can be seen in action on the Hermitage Museum website (
http://www.hermitagemuseum.org/). UK examples include the AHDS Visual Arts' (
http://ahds.ac.uk/visualarts/) and IMAGINE (
http://www.imagine.org.uk/) collections. The wikipedia's entry on CBIR maintains a good, current set of links (
http://en.wikipedia.org/wiki/CBIR).
Some adult-site blockers use simple content analysis to filter out adult images, but these are not always effective, since they are based on screening out images with particular (flesh-coloured) tonal values.
so say I have a photo of a person or an object -- and post it up somewhere on the internet -- no one can take that photo and search and find the exact same photo or that object. Impossible. They might be able to read the source of the image photo hosting site or the type of camera, resolution etc.
I had someone claim the impossibe and I just wanted to confirm it.
Yes, up to now, you are right with your perception of that part of reality.
Yet these people claim that they used google and just typed in a brief description and looked through several pages of images and found a match in "5 minutes". Might just be luck but I know that they must have known where to look in the first place.
It is not quite what you were asking, but there are ways to undo the encoding on a picture that has been edited. There is a guy who has made a website from people's pictures he grabbed from the net and undid their cropping to reveal what had been cropped out.