Handwritten Mathematic Expression Conversion to Docx

This paper aims to embed Handwritten Mathematical Expressions (HME) directly into a Docx document. Writing Mathematical Expressions within a WYSIWYG (what you see is what you get) editor is a cumbersome task which requires a lot of manual effort, which this paper tries to automate. Methods: The task of Recognizing Mathematical Expression is bifurcated into two sub-tasks i.e. structural analysis and symbol recognition. This paper proposes to use deep learning techniques to do these sub-tasks using an end-to-end Densenet based encoder and Attention-Based decoder model, respectively. Findings: The model is trained on CROHME (Competition on Recognition of Online Handwritten Mathematical Expressions) dataset which consists of InkML files. These InkML files are initially processed to generate images and MathML from them. Novelty: We have been successful in creating docx from HME with accuracy trade-off of 1–2% by significantly reducing computational complexity than any other Web application based pre-existing techniques.