An Improved Method of Formula Structural Analysis
Structural analysis is an important step in mathematical formula recognition system. In this paper, an approach of structural analysis based on baseline strategy is proposed in which two strategies are employed to improve the robustness of the system. Fir
- PDF / 265,048 Bytes
- 8 Pages / 430 x 660 pts Page_size
- 42 Downloads / 233 Views
Faculty of Mathematics and Computer, Hebei University, 071002 Baoding, China [email protected] 2 Library, Hebei Universyty, 071002 Baoding, China [email protected]
Abstract. Structural analysis is an important step in mathematical formula recognition system. In this paper, an approach of structural analysis based on baseline strategy is proposed in which two strategies are employed to improve the robustness of the system. Firstly, a converse-matching algorithm is employed to solve the problem in traditional baseline method. Secondly, the feedback mechanism is built to correct the errors coming from the recognition result. The experiments show that the approach can obtain favorable results. Keywords: Formula recognition, Structural analysis, Baseline, Converse Matching.
1 Introduction With the high-speed of science and technology at full speed in our country, the literature of science and technology to which a large amount of mathematical formulas are applied in more and more fields. Current OCR (Optical Character Recognition) system shows high accuracy in recognizing the characters in printed documents, but has no way to handle the mathematical formulas among them. Therefore, it is necessary to develop the method of mathematical formula recognition to extend the application fields of traditional OCR technique. This paper is focused on the structural analysis and comprehension of printed mathematical formulas. The problem has attracted the attention of several earlier workers. Anderson [1] adopted a purely top-down approach for parsing mathematical expressions. Chan and Yeung [2] designed on-line system to recognize mathematical expressions by using of structure and the syntax method. Zanibbi [3] etc. used a transformable technique called tree transform, where the information was represented as an attributed tree. This method consisted of three stages. In the first stage, an original Baseline Structure Tree (BST) was constructed to mainly describe the twodimensional arrangement of input symbols. In the second stage, the Lexical pass produced a Lexed BST from the initial BST by grouping tokens comprised of multiple input symbols. At last, Lexed BST is translated into an Operator Tree, which describes the order and scope of operations in the input expression. But the above-mentioned method is confined to some certain special type of mathematical formula only. In order to resolve the structural analysis problem based on baseline method, this paper proposes a converse matching method with syntax G. Wang et al. (Eds.): RSKT 2008, LNAI 5009, pp. 692–699, 2008. © Springer-Verlag Berlin Heidelberg 2008
An Improved Method of Formula Structural Analysis
693
rules and carries out a post-processing on the mathematical formula analysis. The experiment has indicated this method that can improve the accuracy of the structural analysis effectively. This paper is organized as follows. In Section 2, the concepts of baseline are introduced and some shortcomings of the baseline method are discussed. In section 3, the proposed m
Data Loading...
