Applying hashing search and fuzzy fault-tolerant algorithms for the fast recognition of multi-font printed Chinese characters

In this paper, we apply hashing search and fuzzy fault-tolerant recognition algorithms for multi-font printed Chinese character categories. Every Chinese character category is considered as a graph and represented by the graph description matrix which consists of the types of all its strokes and the spatial relations among strokes. A graph-associative memory is used to contain the pair (graph, content) for which f (graph) = content, where f is the associative mapping function and the content is the associative index of the graph. The associative map ping function f is designed to be a simple arithmetic function by using the hashing technique. Thus, the transformation is performed quickly by mapping the graph to its associative index. In consideration of practical situations where the input characters may be skewed or the thickness of them may be varied when they are acquired from an input device, an attributed relational graph is utilized to represent the Chinese character, and a fuzzy deformation model for the attributed relational graph is developed to construct a fault tolerant algorithm with N modular redundancy. The experimental results show that these techniques make the recognition work well.