Page segmentation and classification using fast feature extraction and connectivity analysis
Jaakko J. Sauvola, Matti Pietikäinen · 2002
Page segmentation and classification are important parts of the document analysis process. The aim is to extract and classify different parts of the page. This paper proposes an approach in which these two phases are combined. The integration process includes fast feature extraction with rule-based classification and label propagation using connectivity analysis providing classified areas in three categories: background, text and picture.