Is SAPIR Vapor?
September 16, 2009
eWeek ran a story that intrigued me. The title: “IBM, European Union Unite on SPIR Multimedia Search Engine”. SAPIR is an acronym for Search in Audio Visual Content Using Peer to Peer Information Retrieval. I like the acronym. It reminded me of Edward Sapir, a noted linguist, and the Latin sapere, which means wise. The question I had was sparked by this passage in the eWeek article:
SAPIR indexes the content of each image and clip using descriptors such as text, color, layout, shapes, or sounds, to help users find comparable images. For example, SAPIR scans a digitized photograph or the bit streams in an MP3 sound file, even if they haven’t been tagged or indexed with descriptive information; this is because SAPIR automatically indexes and ranks multimedia content users upload to the Web for easy retrieval. In a demo, Mass showed how a search on the keyword “dolphin” returned dolphin photos of similar colors and shot angles that users had uploaded to Flickr. After clicking the “similar” link on top of one of the photos, eWEEK saw photos that resembled the dolphin photo in color and shape but did not necessarily include dolphins.
I saw a demo of an image recognition system in Japan when I lectured at Kasai Institute of Technology many moons ago. It worked – sort of. I have also reviewed facial recognition software for a couple of outfits. Great if the person for whom I looked was stationary, looking directly at the CCD, and was not wearing a hat, sporting sunglasses, and distorting his face in a scowl, snarl, or charming grin.
The notion of a demo is very different from a system that has to process images in the wild and return useful matches. My thought is that IBM revs its public relations engine, makes a few phone calls, and sends out suitably rumpled IBM lab wizards. The result is an acronym like SAPIR, but at this point I think SAPIR rhymes with vapor from the Latin root that means steam. Example: the tea pot’s spot spewed vapor from the boiling water.
I recall reading in January 2009 that European researchers “achieved a break through by developing a power image recognition application with mass market appeal.” You can read that story on Science Daily. The technology was MOBVIS and I have lost track of the technology. I think it is easier to issue a news release than to get software to figure out that a photograph is “about” something. For example, I took a picture of Tess and Tyson for this Web log. What is the image “about”? Two dogs or the point that when we sit through a vendor’s demo, the Beyond Search team is ready for a nap. Slippery stuff matching the “image” with the “meaning” in my opinion.
Stephen Arnold, September 16, 2009
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
