Formulas Extraction And Symbols Location In Printing Mathematic Expressions Recognition

The study and compare to the typical systems of math formula recognition currently international have been done in this paper, as well as discuss about the problem emerged when the systems treat with the Chinese document. Based on the existing systems, a new math formula recognition system special for Chinese document is presented. Because the new system makes good use of the characters of the Chinese character, contrary to the existing method, the new system has simplified much more in the structure, at the same time, the efficiency has been improved in a sense. Another advantage of the new system is that it can extract the math formula under the condition that the math symbols haven't been recognized, which is helpful for improving the efficiency of the system. The structure of the new system is very simple, which is constituted of three steps. The first is to image pretreatment, whose aim is to transform a color image to a gray image, to remove the noise, to enhance the image. The second is to extract the math formulas, whose main aim is to separate the math formula form the plain text. This process is constituted of two parts. One is to extract the isolated formulas, the other is to extract the embedded formulas. The last is to locate the math symbol, whose aim is to get the position and the size of the each symbol in the math formulas under the condition that the formulas have been extracted.The system mainly makes preparations for the next process as the structure analysis, symbol recognition etc, which is a difficult in the whole math formula recognition. The accuracy to the extraction of the isolated formulas has arrived at 98%, but the aspect to the extraction ofembedded formulas need improved, at the same time the system has been able to nicety locate the connected components.