Handwritten motives image recognition using polygonal approximation and chain-code

摘要

The present paper proposes a novel algorithm for recognition of handwritten forms. The object of this paper is the pattern recognition of handwritten craft motives images. It is appropriate in one the first time to transform the forms by a polygonal forms using polygonal approximation. In the next step, the extracted the chain-code features, this features as extracted from the contour of polygonal forms. Chain code is a sequence of code directions of a polygon form and connection to a starting point which is often used in image processing "8-neighborhood method has been implemented". Aggregation chain code and the normalized chain code build the extracted feature vector for each image motive. These extracted features are used to train a feed-forward back-propagation neural network employed for performing classification and recognition tasks. Extensive simulation studies show that the recognition system using chain code features provides good recognition accuracy while requiring less time for training.