Speedup of Parsing for Recognition of Online Handwritten Mathematical Expressions

摘要

This paper proposes a method for speeding up parsing process for recognizing online handwritten mathematical expressions (OHME). We prune infeasible partitions in the parsing table to reduce the time for the parsing process. Low score partitions are candidates for pruning. Our method can be applied for any parsing algorithms that use score functions. In this paper, we use a stroke order free system as a baseline system. The method is as follows. First, we analyze the scores of partitions in each row of the parsing table. Then, we determine a threshold for each row to prune low score partitions. Finally, we employ these thresholds to prune low score partitions on the baseline recognition system. The results of evaluations of our method on the CROHME 2014 database show that the recognition process is speeded up by 3.46 times and 4.97 times while recognition rate is reduced only 0.31 point and 0.71 point, respectively.