An important problem that decreases the accuracy in Thai Printed Character Recognition system is the errors in segmentation process. In vertical connected character segmentation, we can easily use the reference lines of Thai language structure. This paper thus proposes a method for detecting the connecting points in horizontal connected characters. First, we extract the features of the connecting points in the character images. Then, we employ Inductive Logic Programming to produce the rules that will be used to classify the unseen examples. Finally, we use Backpropagation Neural Network to make these rules more flexible. The results show that our method achieves 94.94% of accuracy. © Springer-Verlag Berlin Heidelberg 2003.
CITATION STYLE
Pipanmaekaporn, L., & Sachdev, A. (2003). Classification of connecting points in Thai Printed Characters by combining Inductive Logic Programming with Backpropagation Neural Network. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2756, 563–570. https://doi.org/10.1007/978-3-540-45179-2_69
Mendeley helps you to discover research relevant for your work.