2013 | OriginalPaper | Buchkapitel
Edge Based Graph Neural Network to Recognize Semigraph Representation of English Alphabets
verfasst von : R. B. Gnana Jothi, S. M. Meena Rani
Erschienen in: Mining Intelligence and Knowledge Exploration
Verlag: Springer International Publishing
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Graph Neural Network based on edges is introduced in this paper and is used to recognize the English uppercase alphabets treating their corresponding graphs as semigraphs. Graph Neural Network(GNN) is a connectionist model comprising of two feedforward neural networks (FNN) called transition network and output network connected by recurrent architecture according to the graph topology. The characteristics of the edges in a graph are considered as input for the transition network and the stabilized output of the transition network are taken as input for the output network. Edge based GNN is trained using error gradient method. Experimental results show that GNN is able to identify all the 26 graphs of alphabets correctly.