Ebook: Reading and Learning: Adaptive Content Recognition

The amounts of information that are ?ooding people both at the workplace and in private life have increased dramatically in the past ten years. The number of paper documents doubles every four years, and the amount of information stored on all data carriers every six years. New knowledge, however, increases at a considerably lower rate. Possibilities for automatic content recognition in various media and for the processing of documents are therefore becoming more important every day. Especially in economic terms, the e?cient handling of information, i.e., ?- ing the right information at the right time, is an invaluable resource for any enterprise, but it is particularly important for small- and medium-sized ent- prises. The market for document management systems, which in Europe had a volume of approximately 5 billion euros in 2000, will increase considerably over the next few years. The BMBF recognized this development at an early stage. As early as in 1995, it pooled national capabilities in this ?eld in order to support research on the automatic processing of information within the framework of a large collaborative project (READ) involving both industrial companies and research centres. Evaluation of the results led to the conclusion that research work had been successful, and, in a second phase, funding was provided for the colla- rative follow-up project Adaptive READ from 1999 to 2003. The completion of thesetwoimportantlong-termresearchprojectshascontributedsubstantiallyto improving the possibilities of content recognition and processing of handwritten, printed and electronic documents.

---

This book presents major research and development findings achieved in the project Adaptive READ funded by the German Federal Ministry of Education and Research. Beyond techniques and systems for reading and recording documents, the emphasis of this project and the resulting book is rather on the formulation and prototype-style implementation of concepts to make document analysis systems adaptive in real-world scenarios. Furthermore, learning techniques are presented allowing for efficient content recognition and intelligent retrieval of information.