Online handwritten mathematical expression recognition

Widespread usage of smartphones and tablets and their input and output units which are on the same surface, make softwares which support to automatic detection of handwritten symbols and drawings and allow to reorganization in creation of e-documents, more important. Aim of this study is to develop a system which supports to recognition of online handwritten mathematical expression and solving equations and creation of its documents on touchscreen devices. The online handwritten mathematical expression recognition system is composed of five main modules: Symbol detection, symbol recognition, parsing, exporting and solving. In symbol detection, minimum spanning tree data structure is used so as to detect symbols from the strokes of expression. The detected symbols are sent to the symbol recognition module. Here, strokes of the detected symbols are converted to standart form by being preprocessed.Soon after features which is used for classification are extracted. Afterwards, symbols are recognized by using k-means clustering algorithm. All the recognized symbols are sent to parsing module in order to find hierarchical structures of the input expression according to predefined mathematical rules. Then parsed expression is converted to AsciiMathML and MathML which are standard mathematical markup language. Finally, text expressions are sent to solving module so as to determine the types of equations and solving process is performed.