Holistic recognition of handwritten character pairs

Researchers have thus far focused on the recognition of alpha and numeric characters in isolation as well as in context. In this paper we introduce a new genre of problems where the input pattern is taken to be a pair of characters. This adds to the complexity of the classi"cation task. The 10 class digit recognition problem is now transformed into a 100 class problem where the classes are M00,2, 99N. Similarly, the alpha character recognition problem is transformed to a 26]26 class problem, where the classes are MAA,2,ZZN. If lower-case characters are also considered the number of classes increases further. The justi"cation for adding to the complexity of the classi"cation task is described in this paper. There are many applications where the pairs of characters occur naturally as an indivisible unit. Therefore, an approach which recognizes pairs of characters, whether or not they are separable, can lead to superior results. In fact, the holistic method described in this paper outperforms the traditional approaches that are based on segmentation. The correct recognition rate on a set of US state abbreviations and digit pairs, touching in various ways, is above 86%. ( 2000 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.