Document Segmentation And Region Classification Using Multilayer Perceptron
N. Priyadharshini, Vijaya MS · 2013
A document comprises lot of knowledge and documents are considered as the common mode of sharing information to others. Pursuance of information from documents involves lot of human effort, time consuming and can severely restrict the usage of information systems. Thus automatic information pursuance from the document has become a significant issue. It has been shown that document segmentation can help to overcome such issues. Document segmentation is a process of splitting the document into distinct regions. This paper proposes a new approach to segment and classify the document regions as text, image, graphics and table. Document image is segmented into blocks using Run length smearing algorithm and features are extracted from each blocks. Multilayer perceptron, a supervised learning technique has been used to construct the classifier and found 97.49% classification accuracy.