Abstrakt

A NEW APPROACH FOR RECOGNIZING OFFLINE HANDWRITTEN MATHEMATICAL SYMBOLS USING CHARACTER GEOMETRY

Dipak D. Bage , K. P. Adhiya , Sanjay S. Gharde

There are several problems in pattern recognition system like feature extraction problem and identification, pre-processing and classification problem etc. One of the application domains in pattern classification is handwritten character or symbolic recognition. Identifying handwritten characters is always a complex and challenging task for the researchers. Wide research has been done on the character recognition but handwritten mathematical symbol recognition still remain either untouched or remarkably less research has been done. This is also treated as one of the subset of character recognition. Hence the new approach for offline handwritten mathematical symbol recognition system is described throughout this paper. Proposed system is identified by comparative study of feature extraction techniques. For this research, Character Geometry as feature extraction technique and support vector machine as a classifier is proposed. Proposed system can be useful in an embedded as well as mobile application. Also, it can be useful while converting pdf documents into word from where mathematical symbols can be correctly identified. So definitely it will be feasible in long run.