While the primacy of search to the Internet's continued growth and absorption of all other media can't be denied, to date the greatest successes int he search space have been in text, links, and "Google juice." More complicated forms of media, like video, audio and images have presented serious challenges for companies trying to index conent in a searchable way. Incogna, an Ottawa-based firm, has just unveiled a new way of searching images that may propel it past competitors. Instead of searching meta-tags, Incogna searches the image itself, right down to th pixel. Sounds simple, but according to founder Kris Woodbeck (as told to meme.ca), the process is extremely complicated. Complex it may be, but you can test out the results yourself, as Incogna.com has gone live.
The major image search engines on the web today don't even look at the content of their images, they've made themselves in text processing, so they only look at each image's text meta-data. At Incogna, we're trying a different approach. By using highly parallel processors, we process the shape in every image and use this to build a large scale searchable visual index. With typical desktop processors, this scale of... more
blog comments powered by Disqus