Structure-based recognition of handwritten chemical operators

In recent years, recognition of online handwritten chemical formulas becomes a hot area. For the common inorganic chemical formula, locating and recognizing the operators are key issues. This paper summarizes the chemical operators and divides them into three types. Based on the features of these operators, a novel 3-layer recognition model is proposed, which consists of pre-processing, segmentation and substance recognizing. Taking the heating operator and gas operator as examples, we describe the major structural features of chemical operators. Experimental results show that the method is robust and feasible.