IRIS is a service that provides information related to an object just by taking a picture of it. The object needs no special packaging, tag or marker. IRIS uses Telefonica I+D visual object recognition technology. This technology enables to identify objects such as supermarket products, logos, or posters, among others, without any specific tag. Once an object is recognized, IRIS provides a link to related services (website, video, etc.) or information.
The IRIS mobile application works as follows: the mobile user takes a picture of one or more objects for which she wants extra information. The objects are recognized and the application displays a list of identified + related objects. By selecting one of the listed items, the user is directed to specific services (different websites, locations, etc.) associated with that item. Imagine you have an allergy and the specific ingredient is not listed by the packaging. IRIS could provide extra information customizable by the producer that would reassure you as customer.
The IRIS project addresses several problems. First, we want to address an advanced mobile user experience: people obtain extra information just by taking a picture of an object. This is opposed to typing a text query which is cumbersome in a mobile keypad. Second, to link an object to digital data stored elsewhere. IRIS has the opportunity to address any user with a camera-enabled mobile device (which is widely spreads across current mobile devices).
Many visual recognition projects have appeared in the last few years. With the introduction of flat rates for data traffic in mobile devices and the common capacity of taking reasonably high-quality pictures, new opportunities arise for an enhanced interaction with digital content.
As a result, there are already in the international market several Mobile Visual Search applications. Examples are Kooaba or Google Goggles. Kooaba and others of the kind, link the recognition to e-shops such as Amazon. Goggles, returns a Google search related to the query image.
The main difference between those services and our project is that IRIS enables a customisable response by the content provider. In our case, the content provider (e.g. product manufacturer) can choose whether to redirect the user to a website, provide text, locations, videos or multiple combinations of all those options. In our case, even a SME could easily introduce information into the system for a product in its portfolio.
The visual recognition technology developed at Telefónica I+D can also be applied to domains such as that of Augmented Reality opening the opportunity for advanced mobile contextual services. Following this path, we are leading a project called MobiAR in collaboration with five other research centers, companies and public tourism institutions. This Project is funded by the Spanish government under the Avanza I+D program and will lead to a trial with real users by the end of 2010. Our technological development has derived into one patent already filled and another one under way.
IRIS has been awarded as the best innovative application in the BDigital Congress, in Barcelona (May 2010).