Development of Corpus and Deep Learning Model for Handwritten Mathematical Expression Recognition

Abstract

Being the universal language of science, mathematics witnesses its importance by being an newlineessential part of most scientific and technical literature. The scientific literature contains ample newlinenotes of mathematical symbols and expressions that are not easily recognizable as the plain newlinetext. Moreover, mathematical notations are part of visual language. Visual language is defined newlineas a communication system using these visual elements, like graphics. It can be comprehended newlineby its two- or three- dimensional graphics rather than the linear text. Due to its inherent visual newlineelements, the recognition of mathematical symbols and expressions becomes a challenging task. newlineMoreover, the 2-Dimensional structure of the math symbols could disseminate knowledge in newlineall essential technical and scientific literature. Therefore the identification and realization of newlinethese mathematical notations become a natural essence of great practical importance. As the newlinepresence of mathematical text is found in the bulk of scientific and research associated newlinedocuments and literature, searching and accessing math text is of significant importance. Also, newlinedigitizing scientific documents could lead to the enhancement of digital data. Thus, scoping up newlinethe way of easy retrieval. newline