What's Photo Walker? [Japanease]



Walking photographs

PhotoWalker is a temporary name of the web site for virtual walk-through contents created by using a brand-new image technology 'STAMP' (Spatio-Temporal Association with Multiple Photographs). This is developed at the Center for Spatial Information Science (CSIS) of University of Tokyo. We named this site after impression of the technology: you may feel you are walking in photographs or the photographs are walking. If you want to experience this, select one of the contents in World Map and get in the world of PhotoWalker.


Progress of virtual reality

In these years, thanks to technologies, a lot of people are enjoying virtual reality in their computers. Image-based approach is one of the virtual reality technologies, which creates the virtual world by using only photographs. Among them, a well-known method is QuickTime-VR (panorama VR/object VR) of Apple Computer, Inc. In this method, however, there are some problems of the operations to restrict free movement; for example, you may not move into deeper side of photographs (different from zooming in/out). STAMP has solved this problem by allowing vague joint of spaces, so as to move anywhere in photographs. For this technology, we obtained national patent (2001-174582) and international patent (FPA-2093-US).


Expansion of collage

Our technique is also called 'multi-dimensional photo collage'. Photo collage, in art world, is the method of making a picture by sticking pieces of photographs onto a 2-dimensional canvas. If this method is implemented in a computer, pieces of photographs are skewed and stuck to build an interactive pseudo 3-dimensional space. In the same manner as photo collage of fine art, you may build a space by mixing any kind of media. For example, you may mix old and recent photographs, photographs of your face and other's face, or photographs and paintings. If you arrange the technique, this method has great potential to create various contents to represent transformation over time like videos or movies, shift of multiple viewpoints, or selection of the ways to go like a maze.


What is multi-dimension?

Perfection of the technique of photo collage came from Moving Focus series of old master David Hockney, as diversified representation with multiple viewpoints. Even at present, a lot of engineers in the world have been trying various approaches for this technique. Simran Singh Gleason uses full advantages of the photo collage technique in his works, Enfolding Perspectives: Photographic Collage. We may say that, STAMP takes over their techniques of photo collage and materializes multi-dimensional spaces in a computer screen, which are built on individual cognition and are independent of orthogonal three axes. We recommend that you produce new spatial representation of yours by using STAMP.


Using a digital camera...

Digital cameras become widely used recently. Its spreading speed is explosive. We should say that the environment has been completed in which lots of people may collect vast amount of photographs in a short period of time. The quantity (number) and the facileness is beyond the comparison to the past. Along with this expansion, the usage of those photographs has been changing. For example, a photograph is attached to an e-mail or it is released on the Internet; furthermore, appreciating photographs through VJ effects becomes much more popular than ever. Beside such usage, STAMP is useful to arrange photographs in aggregate, not to treat them as separate ones. STAMP enables you to arrange photographs spatially so that you view around the space of those photographs as if you move around in that space. STAMP, in a sense, is the tool for a real-space photo album (creating a photo album from space or creating a space from a photo album).


Browsing pages on the Internet

The radical principle of the Internet is 'hyperlink'. On a normal browser, pages are browsed by tracing their links one by one. If, however, you can browse a subsequent page in semitransparent image on the current page, in other words, if you can view the current page and subsequent pages at the same time, does web browsing become much more convenient? This can be the problem of interface -- how to present database. STAMP is the technology that combines photograph images on the Net with hyperlinks and displays those images, the current photograph (opaque image) and subsequent (semitransparent image), at the same time. This is almost the same as the FOG effect of virtual reality in usual cases.


Introduction of animation

Another feature of STAMP is 'animation'. When you move from a page to another on the Internet, the screen images are changed in a moment. If, to this change, transition representation is applied, how would it be? STAMP introduces morphing technique between pages to represent smooth transition based on human eyes, which is recognized as a suggestion of a new type of browser. Up to the present, the current Internet browser was suggested as an imitation of 'book' and VRML was suggested as an imitation of 'city'. Some user of STAMP pointed out that STAMP is close to characteristics of human memory space. So it turns out that the next subject of STAMP development is transition speed?


Comparison with videos

What is the difference between STAMP and video? This is one of the frequently asked questions. To this question, we always answer as follows: "Video is a recording method provided with time width; photograph, in contrast, collects significant scenes by picking up only impressive moments. Therefore, STAMP has more potential for communicating those scenes with others." Although we answer in this way, STAMP provides the method to use videos, which is in preparation for release. In addition, we are planning to develop various functions: BBS-type STAMP, STAMP operated in association with maps, QuickTime VR, or IBNR. In a series of STAMP projects, we are aiming at associating any kind of data 'spatially' on the Web.


As a record of architecture

Photographs on magazines were the tools in most cases to provide architectural information. This caused photographic architecture to reach cataloging of architecture. There were some opinions that photographs and architecture had complicity in a crime.
Distribution of architectural information via the Internet has just begun. At present, in most cases, photographs are just presented on the Web, while new approaches have been tried; such as editing of architectural information by using virtual 3-dimensional space or technical drawings. We think that STAMP also contributes to communication of architectural information, and it is the best technique to describe 'sequence' that represents transition of view along with traffic lines. It will be interesting if STAMP-like architecture is discovered in the near future, which is one of our motives to push our project ahead.


From the viewpoint of cognitive science

In closing, I have one question here to you: Why STAMP provides views with 3-dimensional depth? This is because the depth is nothing but an illusion since there is no information of 3-dimension in photographs. Then, what fosters illusions in your mind? That is just 'transformation' of images.

Conventional theory of perception had consistently explained perception through the theory of form. Since, however, 'movement' makes information, what turns out to be significant to perception is 'transformation', not 'form'. 'Form', the basis of research of perception, is the unit of geometry. Why has research of perception been sticking to particular geometry? What is viewed through changes of targets is not 'form', but the targets themselves, their real 'shape'. 'Shape' is perceived, not by 'form', but by 'transformation' that has no particular form.
(Masato Sasaki, "Affordance")


Space in a computer, interface, memory of users who view the space, characteristics of photographs... If we seek the essence of those things, we reach the problem of recognition or perception of human beings. STAMP is still incomplete to represent virtual 3-dimensional spaces. Despite this, isn't it possible to say that STAMP is the representation method that extremely takes advantages of human cognitive features by using its nature of automatic compensation to (try to) perceive given information.




Copyrights (c) 2003 hiroya tanaka allrights reserved.