An improved algorithm for recognizing mathematical equations by using machine learning approach and hybrid feature extraction technique

Handwritten mathematical symbols and equations recognition has captured a lot of concentration in the field of pattern recognition. Using efficient multilayer perceptron feed forward back propagation neural network with training algorithm gradient descent with momentum and adaptive learning definitely improve the performance and accuracy of proposed system. By considering hybrid feature in recognition system, the speed is enhanced with tremendous recognition accuracy. An experiment has been carried out for numerous kinds of equations in handwritten form and methodology has exposed successful results. In future, the proposed system might provide key factor to initiate for paperless environment by digitizing and transforming current paper documents.