I came across two interesting experiments involving location metadata and media today.

First, scientists Roberto Cipolla and Duncan Robertson at the University of Cambridge are building a system for inferring location from image content:

Roberto Cipolla and Duncan Robertson have developed a program that can match a photograph of a building to a database of images. The database contains a three-dimensional representation of the real-life street, so the software can work out where the user is standing to within one metre.

Their project is the inverse of the Mobile Media Metadata project, which aims to infer image content from contextual metadata (including location). It is interesting to consider how the technologies might work together: taking a picture of a building tells the system where you are (South Hall). Knowing where you are tells the system what you are doing there (attending class), from which it can infer who the people in the picture are (your classmates)…

Second, artists Pall Thayer, Sara Kolster, and Pete Gomes are playing with the concept of geocinema, using open-source tools to superimpose GPS coordinates on video on real-time. Cool, but how much more interesting would it be if they could:

1. convert those coordinates to higher-level semantic location metadata (”the place I passed out last night”), and
2. use that metadata not just for superimposing on the video but as input for determining the structure of the video narrative?