Everyone knows about OCR (optical character recognition), which allows a computer scanner to "read" a text on a physical document, interprete it and restitute it as a formatted text file. Recent versions include image restitution within a given text, but if the text is indeed gramatically interpreted depending on the language, the image are not. They are restituted as they are, bits of color information on a two dimension surface. Which is what they are, as far as optics are concerned.
But this is all about to change, and sooner than we thought. The BBC announced today that a research team has created a prototype software able to actually interprete an image or movie. In other words, an optical recognition of objects on a multimedia document, coupled with a robust artificial intelligence system. The prototype is limited to certain types of scenes, but only raw power prevents the technology to be applied to anything, from your holiday pics to movie rush analysis...
This is going to revolution multimedia software. Imagine for instance a software like Picasa, but with the option of automatically adding relevant caption to each image, sorting them in a database in a couple of minutes, based on the actual content and topic of these images, without any human intervention... And think also, this is not only going to revolution your holiday pics, it is going to revolution intelligence work. Big time!