IBM, European Researchers Building Multimedia Search Engine
IBM and a consortium funded by the European Union have developed a multimedia search engine for untagged content. They claim their results are better than that produced by Google and Yahoo, according to an IDG story in The San Francisco Chronicle.
Called SAPIR, for Search in Audio-Visual Content Using Peer-to-Peer Information Retrieval, it analyzes “low-level descriptors,” such as color, layout, shape and sounds, in photos and videos and compares them to others.
As ReadWriteWeb explains it, you could submit a photo, perhaps of a locale you can’t quite remember, and the search engine could pull from a huge database of similar photos and identify the location. Or you could submit a sound on an audio file and have it identified.
So far, this research is in its early stages.