The Research Of On-line Handwritten Mathematical Formula Recognition System

Abstract:A handwritten mathematical formula recognition system should be divided into two major steps: the character segmentation and structure analysis. Among them, the character segmentation is the basis of the formula recognition. The segmentation will have a direct decision of the correct rate, so a good segmentation algorithm would be more important. The structural analysis is the key of mathematical formula recognition. However, the mathematical formula recognition process very difficult even when all the individual characters and symbols can be recognized correctly.In the last years, research in the field of pattern recognition, handwritten mathematical formula recognition has been more and more popular. However, the diversity of approaches to the problem and the lack of a commercially viable system, which show that there is still much research to be done in this area. In this paper, I will describe an on-line handwritten mathematical formula recognition system based on the formula structure and two-dimensional grammars. In general, this approach can be successfully used to analyze the complex structure of relationships. An important benefit of the structural construction is in not treating symbols segmentation and structural analysis as two separate processes which allows the system to perform segmentation in the context of the whole formula structure, helping to solve arising ambiguities more reliably. We explore the complexity of polynomial algorithm and extend the grammar by many new grammar production rules, so that the system became more effective for formula. We propose several grammar extensions to support a truly wide range of mathematical formulas, and in the system to implement the new features. Our current method can recognize functions, limits, derivatives, binomial coefficients, complex number and so on.In paper, Chapter I describes the research background of the online handwritten mathematical formula recognition, research status, some of the commercial product and the work done in this paper.Chapter II describes the relevant methods of symbols preprocessing and detection symbols based on OCR tools. The user handwriting strokes are grouped, based on different grouping symbols to get some candidates, combined with the two-dimensional structure of mathematical formula and related the context information to determine what stroke group is the right kind of group, which should be a meaningful symbol.Chapter III focus on the general steps and common methods of the on-line handwritten mathematical formula recognition, and the improvement in this paper. Chapter IV introduces the major tasks of the handwritten mathematical formula recognition, proposed 2D co-ordinal grammar and parsing algorithm for handwritten mathematical formula recognition.Chapter V focus on the symbols characters of the handwritten mathematical formulas, proposed extension of two-dimensional grammar combined with these features, as well as two-dimensional structure of the formula. Mathematical formula Word EQ domain. Mainly introduces how to design and implement the handwritten mathematical formula recognition system, put forward my own ideas and programmed to complete it.Chapter VI introduces the 2D co-ordinal grammar in this paper used to recognize online handwritten mathematical formulas, illustrate the effect of the system. Introduces time complexity of the system, then put forward the work at the next phase according to the current problems.