Automatic performance evaluation of mathematical expression recognition
Jian-Ming Jin
Hong-Ying Jiang
Kai Wang
Qing-Ren Wang
Inst. of Machine Intelligence, Nankai Univ., Tianjin, China;
Abstract
The recognition of mathematical expression is very difficult. Most recognition methods are still tentative. In order to compare and improve them, it is necessary to evaluate their performances. However, the research on performance evaluation of mathematical expression recognition is very scarce. A triplet tree with marked edge is presented to represent the 2-D structure of mathematical expression. An automatic performance evaluation method, named BUTD (bottom-up and top-down), which is based on tree matching, is proposed. BUTD has been used during the research of an expression image recognition system. Also, BUTD is the only known method, which can find any errors of character segmentation, character recognition and expression analysis.
Index
Terms
Available to subscribers and IEEE members.
References
Available to subscribers and IEEE members.
Citing Documents
Available to subscribers and IEEE members.
Cart
