There is a lot of talk about the Semantic Web and how the knowledge embedded in the data on the net can be analysed to give meaning.This post considers just one aspect of that, in relation to images.
One aspect of this was highlighted through the combination of tagging and software. Photographs tagged with identity, or geographic information as well as temporal markers allows software to combine images from disparate sources and create seamless views of the subject matter from multiple sources. In the TED Talk below Blaise Aguera y Arcas demonstrates how this is achieving the most amazing results.
Alternative Link here >>
The key here is that the tags provide meta data about the picture which can then be interpreted. But, so what? Suppose you add a picture of New York, and tag it, the meta data then can associate photos to yours and you can reconstruct an image that extends beyond the edges of your photograph. It can be linked to aerial and satellite images so that you can zoom in to your album from a world view, and where others have taken higher resolution/closer pictures of your subject you can even zoom in to detail your picture didn’t record. Google maps and Google Earth have used elements of this for a while, changing from sattellite to high level and hten low level aerial photographs as you change resolution. Seadragon makes those transitions seamless.
Considering where this might develop, the temporal data included with photographs means there's the possibility to build time-lapse style movie development into the series. Think about how pollution could be monitored by the rate of colour change of a building, or its degradation by acid rain and so on. Photographs only exist for a relatively short period though, and have been common place only recently. so these things may only be relevant where the changes are relevant to the timescale available. That's really just scratching the surface. As new people add images of the same place these tools continually enhance and add to the knowledge base because of the context added by others in tags and similar meta data items. Like wikipedia errors will tend to be removed as the mass of data centres around the reality.
Pictures are of course just one form of data, video is a series of still images, so they too can be integrated into this sort of knowledge base too. The possibilities there are significant. These technologies are barely conceived yet they are already changing the web.
Now, imagine the collective data that stored as text and applying the same techniques leads to enormous opportunity to expand and enhance our contributions through the semantic capabilities now being added to the web.
I've no doubt that we'll be discussing this sort of development many times here in future.
Twitter this... Click here >>
