|
ViPR represents the state of the art in the field of visual pattern recognition.
ViPR looks for common features in one image, against a database of already known images. Its performance is not affected by uneven lighting, partial object views, changes in viewpoint, and camera-optics distortions. And ViPR does this in an extremely computationally effective way.
This enables users to look up information for images they capture in real-world settings like a bookstore or street corner.
Take for example Amazon's web site: their database contains millions of images of their products. As well, A9 has announced a database with 20 million pictures of store fronts in 13 large metropolitan areas. All these images can be added to a ViPR database.
The image database and processing are done on a dedicated server. Client code does not use up processing power or local flash memory. If MMS messaging is used to send the image to the server, no client code is required, making the deployment and support across multiple platforms a trivial task.
|