Today, a profile on a new product from Toronto-based startup Idée called TinEye. TinEye is a image-based search engine that has just been released to public beta, and its claim to fame is that it can search the Internet for not only exact copies of the image but also derivations of the image in question - the image can be resized, cropped, warped, and manipulated with Photoshop, but the goal of TinEye is to find all of Web pages that use some form of the original image. While the number of pages currently indexed by TinEye represents just a small fraction of the total number of pages out there (currently TinEye indexes 701M pages), the goal obviously is to expand this to a much greater number, and likely with the intention of indexing fresh pages as quickly as possible.

Some obvious uses have already emerged for this form of image search, with clients such as Associated Press, Agence France–Presse, Getty Images, European Pressphoto Agency, and Masterfile using the service to identify when their images are used for the purposes of royalty verification. But the number of potential applications for this technology could extend far beyond what it’s being used for currently.

1) Law enforcement applications of the technology could be pretty exciting. TinEye could be used to find not only criminals but also lost children by searching not only news sites but also casual photography sites such as Flickr.

2) Mobile search is one emerging space that will definitely heat up as the penetration of smartphones rapidly increases. While Google is already leading the market in text-based search, the mobile world brings several important new forms of input that have yet to be fully leveraged: voice and video (via the smartphone’s camera). While voice recognition software will simply reduce this form of input to just text, the problem arises with what to do with the pictures captured by smartphones. Imagine using the smartphone’s camera to snap a vintage car that is stopped at a traffic light, and then using that image with TinEye to immediately identify the exact year and model of the car. If that same search were to link automatically to an eBay auction for the vehicle, the revenue implications for the platform that drove that buying decision could be significant.

3) Asian language search - The big problem with foreigners in reading picture-based languages such as Chinese is that searching in dictionaries for the meaning of the characters is usually slow and error-prone. Pen-based input systems can help, but using a smartphone’s camera to quickly take a snap of it, then sending the picture to TinEye to do an instant translation would be a valuable service for travellers. In a similar manner, travellers looking to get more background on a famous tourist attraction might in the future simply take a picture of it, and expect Mobile TinEye to pull up reviews or the historical background of the monument.

4) Image tagging of a huge catalog of images and videos is an untapped gold mine for sites with large repositories. The ability to display relevant advertising when these media files are accessed is something that Google would love to do with its YouTube product. TinEye’s technology could be used to identify company logos or products within these files, enabling ads to be displayed at the right times.

The bottom line is that this technology in my view has a lot more application that simply royalty-verification for photos. With a whole new generation of users now using smartphones as adoption moves beyond just early adopters, similar efforts by Google, Yahoo, and Microsoft in the area of image-based searching should soon emerge.