The “age of digitalization” is getting to it’s peak. More and more firms as well as individuals are tending to save as much as possible digitally. Most information (mails, pictures, songs, movies, …) is created digitally anyway and the rest is converted (e.g. scanned). All these tendencies just address problems of storage. But at least of equal importance is the search process. What is are all the yotabytes of data worth if we can’t find the right thing at the right moment?
The significance of the “search problem” is been represented by the standing of companies like Google or Yahoo and currently a lot of research and development is done on this topic (e.g. semantic search). Still one of the greatest challenges is finding information stored in image data. In this area two interesting developments are in progress:

• a new image-search method developed at the University of California, San Diego
• the open-source project Ocrupus that aims to develop a high-level, easy-to-use handwriting recognition system

Credit: Gustavo Carneiro, Antoni B. Chan, Pedro J. Moreno, and Nuno Vasconcelos via Technology Review

Imagine what a radical enhancement of the search process it would be if you enter your name into your search engine and then … besides all the well known text results you’ll find pictures with your face on it, with things that are created by you or which belong to you as well as a scanned version of your signature for example. A list of further examples is to be continued indefinitely.

Posted by maz on Sunday, April 15th, 2007 at 21:25.