- Toward Integrity of what you see and what it is-
What if your cell phone could read ordinary objects as if the barcode was put on everyday objects. so that your mobile device could understanding the social context? Familiar Street is an unofficial name for the research I have done in the Interaction Laboratory of Sony CSL(Computer Science Laboratory) with Professor Jun Rekimoto. Detailed information about this research can not be provided on the website yet due to security of the content at Sony CSL.
In prehistoric era, the objects people encounter in their daily lives mostly had been natural objects such as trees, mountains, rivers, and caves. People could understand the nature of their surroundings through intuitive experience because natural objects represent their nature by shapes, sounds, and smells. On the other hand, buildings we encounter everyday in streets do not properly represent its nature through their shapes. They include information that is in huge volume and ití»s too capsulated inside. Due to this, people in modern era are forced to receive visual information that lacks understandable representation, which causes people to feel isolated and surrounded by unfamiliar strangers.
< click to view the video(320x240) >
(quick time player needed to view)
In the familiar street project, we aim to build a mobile device that is familiar with street and building; that recognizes building and artificial structures and makes their information overlapped on the images instantaneously. To achieve this goal, we decided to use "fern", a planar object detection algorithm which detects the objects based on its 2D images. We envision this system being used with large volume of picture database of streets such as Google street view, and this would enable a more intuitive access to our surroundings.
With this street recongizing technology, we would be able to have direct access to "what we see". This project is , in that sense, an extension of the significant paradigm WYSIWYG (What You See Is What You Get). This project seeks to expand on the ideas first developed using the WYSIWYG paradigm. Now, the thing you can get by seeing is not just stuffs on your monitor, but also your surroundings.
<<previous :: next>>