Clay's Blog

Technology Review has an article about a new tool called Zoetrope developed at Adobe’s Advanced Technologies Lab by Mira Dontcheva. The article includes a video of Zoetrope in action.

Zoetrope gives the user access to a range of data interaction tools that harness snapshots of a web page over time. The user can use the DOM to interact with individual components of a page, especially data driven components. “Lenses” can be place over these data-driven areas and the data can be seen over a temporal period. These changes can be graphed or visualized using Zoetrope and can be linked with other lenses on the same page or even other sites. This is truly incredible software. The web is really becoming a giant database and we are reaching a point where tools are popping up all over the place to harness this data and visualize it so that it becomes meaningful in everyday contexts.

Just ran across some amazing work being done by Dan Goldman, who did his doctoral work at University of Washington and is currently working as a Senior Research Scientist at Adobe in Seattle.

The research is focused on interacting with video and with objects in video and relies on current work in computer vision. The technology allows users to interact in some really amazing ways with video for annotation and motion analysis. The process uses a storyboard metaphor to visualize a short video clip in a static image. The user can manipulate spatial relationships in the storyboard image in a natural way to interact with the video stream. Some details and references are available on the Adobe Technology Labs site. Check out the technology in action in this amazing video clip on Vimeo:

Dan Goldman – Interactive Video Object Manipulation Project

Saw an interesting post on TechCrunch yesterday. YouTube recently rolled out a new feature that is a welcome addition to their toolbox: deep linking to a point inside a video stream. It’s a very easy implementation, as well. All you need to do is add a ‘#’ at the end of the YouTube URL and then reference the time code following the ‘#’ sign. For example:

http://www.youtube.com/watch?v=1flVlL4Mf8k#t=0m20s

Very useful!

Sony just announced the latest version of the Portable Reader System, the PRS-700. Hop on over to Gearlog for a quick show and tell.

The new version is $400 and includes some nice features. Sony has added a touchscreen for page turns, including turning multiple pages quickly by swiping and holding. A welcome addition is a set of side LED lights for reading in the dark. Sony also announced changes to the online bookstore, which currently truly sucks. I doubt any changes could make it worse.

For complete specs, pics, and complete specifications checkout Sony’s site.

[Edit]

TeleRead has some additional information from the announcement press conference. One thing Paul Biba mentions in the TeleRead post is that this new version is a great deal faster than previous versions and faster than any current competitor. This is apparently due to Sony’s expertise in writing custom drivers and designing the display processor. A faster eReader. Now I really want one of these….

Microsoft Research recently released a great little panorama image stitching utility. You can check it out at the Microsoft ICE project site. The utility is a free download.

One of the really nice features of this tool is that it can export to many different image formats. Once exported, one could bring the image into, for example, a video editing package to do pan and zoom effects for video. In addition, there is an export option for Deep Zoom Tileset that creates a series of stitched images and some XML data that allows the image to playback on the web inside of Microsoft’s SliverLight 2 browser plugin. The result is a nice pan and zoom image similar to what one gets with a QuickTime VR movie. You might have seen this in Microsoft’s PhotoSynth tool. And this is all free. Grab the software and have some fun!

I’m hoping to get a couple of experiments up soon, but I’m waiting on a server configuration change for the SilverLight files to run correctly in the browser. I’ll post them when that happens.

For a fantastic overview of this year’s SIGGRAPH, head on over to Hack a Day and read Eliot Phillips’ post there. Waaaay better than my weak efforts…heheh.

As Eliot points out in his post, most of the papers are online at various locations on the Interwebs and the links are all aggregated at Ke-Sen Huang’s site. Make sure you check out “Finding Paths through the World’s Photos“, especially if you’ve been following Microsoft Labs’ Seadragon implementation Photosynth. Extremely cool.

Well, another SIGGRAPH is history. It’s been a terrific creative battery recharge. Thursday and Friday highlights include a really cool Production Session on how the various visual effects companies that made Speed Racer went about replicating the look and feel of anime in a live action motion picture, a very entertaining and interesting overview on the use of Adobe Flash for animation, and an absolutely fascinating class on transportation visualization.
Read more »

It’s been an inspiring conference so far. The classes I’ve attended have been excellent. On Monday I attended the half-day course on projectors and spatial augmented reality for (I think) the 4th year running. Ramesh Raskar and Oliver Bimber were fantastic as usual. They were joined this year by Aditi Majumber who spoke about large-format displays and Hendrik Lensch who spoke on computational illumination for 3D scene modeling. One of the things I really get excited about in this class is what Raskar calls RFIG. In essence, this entails adding a photosensor to an RFID tag and then projecting structured light from a handheld projector on the photosensor in order to acquire a relative position for the tagged item. With the unique identifier and the relative position, we can query a database and then project useful information about the identified items directly on the items themselves using our handheld projector. All this is made possible by very small and relatively inexpensive handheld computers with wireless network access and attached projectors. You can check out their work, including the full-text of their book, Spatial Augmented Reality, on the supporting website: SpatialAR.com. Great stuff.
Read more »

I arrived at the Los Angeles Convention Center today and picked up my credentials for this year’s SIGGRAPH Conference. If you’re unfamiliar with the organization or the conference, it’s a part of the Association for Computing Machinery; SIGGRAPH is the largest SIG (Special Interest Group) in the ACM. The full name is the Association for Computing Machinery Special Interest Group on Graphics and Interactive Techniques, which is more than a mouthful.This year’s conference has added a couple of new features. The fantastic Animation Festival has been expanded and more deeply integrated with the rest of the conference. Screenings are being paired with artist and tech talks and will be running throughout the day, with special big-studio presentations in the evenings, including Pixar and Dreamworks.

In addition, conference organizers are implementing a new RFID tracking component to gather demographic data about attendance at conference sessions and events. The RFID tags are embedded in attendee ID cards (attendees can opt-out by recycling the card at the conference). This year is a pilot of the technology with expanded use next year if all goes well.

I’m especially interested in how this works out. The ILC is partnering with Tulane’s Middle American Research Institute on some interesting RFID projects as MARI moves to new and renovated storage and display space over the next 18 months. On a related note, the half-day class on computer graphic projector technology, including Spatial Augmented Reality, is in the morning at 8:30. I’m hoping to come away with some good ideas for implementing this technology in our MARI-related projects in the near future. I have a meeting set up on Thursday afternoon with engineers at Motorola on current RFID technology, thanks to the gracious assistance of Ian Thomas, Vice-President of Business Development at O’Neill Software, Inc.

I don’t think I could imagine a better week of total geek heaven.

Just heard about a really cool flickr visualization called Tag Galaxy (Thanks, Mike!). When you visit the site you can search flickr for a particular tag and then build a stack of related tags to narrow a search. Related tags are displayed in a 3D visualization in the style of a planet with each related tag displayed as a smaller satellite object orbiting the original “planet.” As you click through each additional tag in the stack, the process repeats narrowing the search. When you are finished narrowing your search, you simply click the central “planet.” Images meeting your search are arranged as the outer surface of the sphere. You can click and drag to rotate the sphere and click on individual images to pull them out and show them in front of the rest on the sphere. Another click brings the full image up and provides some of the flickr metadata and a link to the flickr page for the image.

As a visualization that provides an easy way to browse a large number of related images quickly, this tool is very successful. Often, 3D interfaces do not provide the most efficient means to accessing data. But in this case, there’s and excellent fit. Take a moment to check out the site. It’s definitely worth the bookmark.

http://taggalaxy.de/