“Riding the Net” by Sommerer

Riding the Net presents a novel approach to browsing the Internet in a more intuitive, playful, and entertaining fashion. While two users talk and communicate with each other, keywords of their communication are picked up by the system’s speech-recognition engine. These keywords are then used to search and download corresponding images from the Internet. When users, for example, speak about “houses” or “flowers,” different images of “houses” or “flowers” are downloaded. As there is usually a vast amount of images available for each keyword, users see new image icons constantly retrieved from the Internet. All images are then collectively displayed in 3D in the system’s interactive window and streamed from the respective view of each user. As images come from either the left or right side of the screen, they all stream toward each other before they leave the screen and are replaced by new images derived from new keywords spoken by the two users. The entire image scenario on the window surface constantly changes, since it is a direct interpretation of the users’ dialogue and communication with each other.

Both users can also touch the image icons on the screen: this halts the images temporarily so users can look at specific image icons in more detail. When they do this, the exact URLs for these specific image icons can be downloaded onto a separate computer screen, so users can find out where the images came from and what they refer to.

Riding the Net provides an entertaining and playful way to browse the Internet, and users become intensively engaged in the vast amount of visual information available from and presented by the system. Users can control the content of what they are watching through their own decisions, dialogue, and interaction.

Technical Description
• Three Pentium III PCs including NVIDIA GeForce2 high-speed graphic cards
• One Internet 100/10baseT hub
• One LCD projection unit: high-resolution LCD projector with true 1280 x 1024 resolution
• One window-detection frame, including IR sensors
• One window-detection interface
• One space construction (approx. 3 x 3 meters), including window glass surface and two chairs
• One graphic software system
• One interface software system
• One speech-recognition and image-retrieval software system

Riding the Net
Christa Sommerer, Laurent Mignonneau, and Roberto Lopez-Gulliver Interactive
Web-based image browser developed at ATR MIC Labs. Interface design support:
Stephen Jones.