Partitioning and searching dictionary for correction of optically read Devanagari character strings

This paper describes a correction method for optically read Devanagari character strings which uses a partitioned word dictionary. The word dictionary is partitioned in order to reduce the search space besides preventing a forced match to the incorrect word. The envelop information of words consisting of the number of top, lower, core modifiers along with the number of core characters form the second level partitioning feature for short word partitions. The remaining words are further partitioned using tags. A tag is a string of fixed length associated with each partition. The search process uses a distance matrix for assigning a penalty for a mismatch. An improvement of approximately 20% in the recognition performance is obtained