The Informedia Project

The Informedia Project is a research initiative to train computers to “read” digital video. Working through Carnegie Mellon, the idea behind the project is that, since video is such an important historical and artistic medium, there should be measures in place for machines to be able to automatically transcribe, search, and index video. In theory, this would make searching for bits of video as easy as searching for text via Google. This is a very cool and technically very complicated task. The Informedia Project Website documents the work being done.

Ironically, for a project that wants to make information retrieval considerably easier, the Informedia Project Website itself is not the most user-friendly. On first look, the site appears clean, straightforward, and easy to navigate. However, when one starts digging around, it becomes apparent that there might be a lack of vision to the site.


The site documents a collection of research in the left-hand navigation bar. Each of these research projects has a summary of the project; many also include a collection of papers documenting the research done for the project. So far so good.


But the order of the nav bar is strange. Projects are listed first, then past work. Following this is Publications, Demos, the Team listing, What’s New, a link to the site map, and finally a link to the home page. The order of the links seems odd enough, but what is most frustrating is that the list of work/publications seems to be split between the project heading and the publications heading. There doesn’t seem to be any logic to this, except that somewhere around 2004 it was decided from that point on to list all work on the Publications page. Within the publications themselves, there is a lot of fantastic information about the work being done within this project.

The site has a Google-powered search bar, which works as one would expect. The amount of information varies from paper to paper. Some papers have links to the full-text; others do not. I would imagine this is due to publication rights.

There is a line on the nav bar called Demos/Downloads. What is potentially a large chunk of an explanation for the site is currently under construction. However, there is a timeline of the various projects. This is pretty cool. It is dynamic and takes advantage of web 2.0. Everyone who designs a website with any substantial amount of information should take note of this. It’s a simple timeline that is much more usable because of its dynamic attributes.


Altogether, the Informedia Project is one that I personally think holds great potential, but for the moment has a website that is underwhelming. Still, I recommend the site for anyone interested in visual search engines.