The classification of online handwriting samples can be effectively addressed by a granular computing approach. In fact, handwriting can be viewed as a sequence of information granules consisting in single strokes. In this paper, an automatic handwriting recognition system is proposed. An oriented sequence of nodes, as a particular directed labeled graph, is used to represent each handwritten pattern. Each node of the graph stores the feature vector describing a single stroke, while the edge connecting each node to the succeeding one stores information about the pen displacement between the two strokes (usually referred as virtual stroke). Once the handwritten patterns have been represented by labeled graphs, a general technique for automatic graph classification is used to perform different recognition tasks. The tackled tasks include word recognition, writer recognition and character set recognition. The tests have been carried out using real world data. © 2007 IEEE.
Online Handwriting Recognition by the Symbolic Histograms Approach / DEL VESCOVO, Guido; Rizzi, Antonello. - STAMPA. - (2007), pp. 686-690. (Intervento presentato al convegno IEEE International Conference on Granular Computing 2007 tenutosi a San Jose; United States nel November 2-4 2007) [10.1109/GRC.2007.4403187].
Online Handwriting Recognition by the Symbolic Histograms Approach
DEL VESCOVO, Guido;RIZZI, Antonello
2007
Abstract
The classification of online handwriting samples can be effectively addressed by a granular computing approach. In fact, handwriting can be viewed as a sequence of information granules consisting in single strokes. In this paper, an automatic handwriting recognition system is proposed. An oriented sequence of nodes, as a particular directed labeled graph, is used to represent each handwritten pattern. Each node of the graph stores the feature vector describing a single stroke, while the edge connecting each node to the succeeding one stores information about the pen displacement between the two strokes (usually referred as virtual stroke). Once the handwritten patterns have been represented by labeled graphs, a general technique for automatic graph classification is used to perform different recognition tasks. The tackled tasks include word recognition, writer recognition and character set recognition. The tests have been carried out using real world data. © 2007 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
