Making Documents Work: Challenges for Document Understanding
ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 2
Hi-index | 0.00 |
Form classification and labeling is an important problem in the context of document processing. This paper presents a new method for exemplary learning the appearance of documents, i.e. empty forms. Taking scanned form images, the system identifies significant layout features and based on that, determines layout reference patterns. These patterns are used to classify and label filled instances of the represented forms. Thorough experimental tests show the reliability of the approach when applied to everyday documents.