As I sit here shivering in the North of England cold and fog, it seems kind of appropriate to be writing about a technology that has been developed by Polar Rose who are based in the even colder region of Malmo in Sweden.
Using face recognition technology originally researched at the Universities of Lund and Malmo, Polar Rose have combined this with a 3D modeling process that creates the 3D image from a single 2D photograph using 2D to 3D polygon extrusion, to produce a unique visual search application. An open Beta phase will start early in the new year, taking the form of a free web browser client plug-in and featuring a royalty free API for partner integration, and should seamlessly enhance the people browsing experience.
Here’s the problem that Polar Rose has solved: currently if you do an image search in any of the available online engines for, say, John Smith then you will get a massive assortment of images including many John Smiths but probably even more which are not people at all but simple images that were located close to text saying John Smith. Polar Rose knows what a face looks like and will filter out everything that isn’t a John Smith so the results contain facial images only, and it will then group the individual John Smiths using that facial recognition technology. It is this ability to perform an automatic search for, as well as recognize, faces on a consistent basis in much the same way that the human eye and brain can do, that sets Polar Rose apart. Everything from variations in lighting, facial emotions and the particular pose used are compensated for in software, which radically improves the final photo matching. The applications when it comes to discovering image based information, especially within the social networking and web based photo album spheres, should not be underestimated.
Indeed, Polar Rose has taken the social networking concept and built it right into the technology. The system allows for ‘collective social intelligence’ so that human refinement of the computer-generated face recognition is possible. Polar Rose’s core technology combined with the training and input by users empowers unprecedented accuracy, speed and functionality. “We are very excited by the enhanced information and linking offered by visionary search, and are looking forward to creating a whole new experience by making photos searchable in a new way”, said Nikolaj Nyholm, CEO of Polar Rose.
Here are just some of things you will be able to do:
- Search for more photos of the same person on specific sites or across the whole Internet
- Collectively add information and tag people in online photos
- Automatically sort their online photos by the people appearing in them
- Be alerted when new photos matching visual search criteria appear
- For an online photo album, to categorize hundreds or thousands of photos automatically by face, as well as finding local or global matches and searching/sharing feeds from friends.
- For a conference the viewer will be able to get or put information on the people appearing on the photo documentation site of the conference. This will help anybody finding a name to a face or vice versa.