Our first application is called "eyeDentify". It is a demonstration of color based object recognition technology for Android mobile phones, as well as a demonstration of Ibis, a distributed programming environment developed at the Vrije Universiteit.

With the eyeDentify application users can teach their phone to recognize objects in photos taken wherever they are. Users begin in "learn" mode by taking a photograph and typing in a short description of the subject of the photograph. Users can then change to "recognize'" mode and take another photo at which point the application will search the database of learned objects for a matching photo and present that photo and a description of the photo to the user.

We believe this application demonstrates the power of "SmartPhone" platforms such as Android and offers many interesting possibilities for future applications of high utility. While more of a fun application than one with great utility, eyeDentify opens the door to several other advanced applications, including a museum docent application, a photo scavenger hunt application, and a buyers guide based on photographs. "Shazam" is an audio recognition system which is able to identify the name and artist of a song based on a sample of the audio taken with a mobile device. With the technology we present with eyeDentify we are able to provide information on an object based on a photograph of that object, thus we offer image recognition similar to how Shazam offers audio recognition. Furthermore, while application offer "buyers guides" based on a product name, our technology could be used as a buyers guide based on a photograph of what the buyer is looking for.

In a museum docent application, users could be given a phone which they use to take photos of paintings or other works of art in a museum and then be offered localized text with information about that work of art. Many museums do not allow photography in part to prevent users from using a flash that may be damaging to the art as well as control the works in their collection. By offering visitors a device without a flash they can control the quality of the photo as well as prevent damage to works because of flash usage. This would benefit both visitors, who enjoy having photographic keep sakes, and the museum, which could charge for the download of photos and control the quality of the images leaving the museum at the same time.

In a scavenger hunt application the organizer could take a series of photos. The players of the game would be required to take photos which the phone identifies as the same object. In a "Shazam" like service users could photograph a product that they are interested in purchasing and then find out where similar products are being sold.

We offer, two versions of our application. The first is a purely local version of the application which does all image analysis on the phone. Because of the resource constraints (available memory and processing power) the image size that the phone can analyze and still remain responsive to users is downsized and the amount of calculation done is reduced. This has a direct impact on the accuracy of the image recognition that can be performed. However, because the Interdroid project is primarily interested in distributed applications, we also offer another version which can offload the computation to more powerful computational resources. This is accomplished using the "Ibis" platform for distributed computing developed at the Vrije Universiteit. We have already run this application on a system of computers spanning the globe in order to provide soft real time recognition performance on high resolution images. Thus the distributed version of the eyeDentify application offers users very large computational resource with which to do the object identification.

Future work for this application will be to enable sharing of the learned images between users using other components from our research from the Interdroid project. In this way we can enable users to learn objects for identification from their friends. This is just one example of the kind of distributed applications we are aiming to support with the Interdroid project. We are also planning to build other distributed applications as part of our work such as ones for agreeing on a restaurant to visit amongst a group of people with SmartPhones, social gaming applications such as poker, and even disaster management applications such as one which can search for a person with a particular skill such as CPR, or a "health and wellness" application which can report the status of friends and family. The goal of the Interdroid project is to make programming these sorts of applications extremely easy, while also providing a high level of robustness by operating in a completely distributed fashion without the use of centralized components.

Please note that, while the application can be run in the Android Emulator, the emulator always returns the same image which is not very interesting. For best results please test the application on a real phone. Also note that our application is available in both English and in Dutch.

As always, source code for this project is available on our gforge machine.

Screen shots from this application can be seen in the Screen Shot Gallery For This Application.

You can download a compiled version of the standalone version of our application here: The attachment id given is not valid. or on our gforge site.