Recognition of Math Expressions & Symbols using Machine Learning

Abstract

The symbols and expressions used in math are very important and used in daily routines. It is necessary to have them in electronic form to access and produce them easily. The goal of this research is to provide a bridge between the knowledge of the organizer and the user inputs. By making math symbols and expressions visible to the non-native speakers, this research could introduce them to math notation. The method is to be recognized mathematical formula as well as symbols (printed or handwritten) based on feed forward back propagation neural network, support vector machine and K- nearest neighbor classifier. The noise free clean & clear input image can be obtained in preprocessing. The elements associated with math equations can be isolated with the utilization of segmentation. Finally feed forward back propagation neural network, support vector machine and K- nearest neighbor classifier with extracting static and complex features is used to recognize the expressions and symbols on the handwritten and printed image too. The receiving operating characteristics (ROC), confusion matrix & overall recognition system determines the accuracy and efficiency of the proposed system. The recognition of handwritten mathematical equations, symbols, digits provides a lot of real time applications as well as non-real time applications and most important application of math equation recognition system is math talk system for visually impaired people.