I am quite sure that the internet is full of recurring images. How often have you seen this little fellow in some Office doc, web page and so on. So why not write a little web crawler that finds these images and then indexes them. This would be quite useful if you have a copyrighted image for example and you want to find out who is using it. Or just to see how many people are using your clip art. If you scale all the images to the same size and allow some minor differences you should be able to write a filter that does this quite easily. Of course you would have to filter all the different images out of ppt (Power Point something) files and web pages. But all this has been done somewhere before.
2 comments:
This was Wills idea from year 1 or 2?
I don't know if you have seen this or if this is what you meant but there is a website http://tineye.com/ which you can search for photo's and it works for ones that are quite radically altered.
Post a Comment