Handwritten Digit Recognition using various Neural Network Approaches

Abstract: Handwritten digit recognition is one of the important problems in computer vision these days. There is a great interest in this field because of many potential applications, most importantly where large number of documents must be dealed such as post mail sorting, bank cheque analysis, handwritten form processing etc. So a system should be designed in such a way that it is capable of reading handwritten digits and provide appropriate results. This paper presents a survey on various neural network approaches to recognize handwritten digits.

Keywords: Artificial Neural Network (ANN), Handwritten Digit Recognition, Back-propagation (BP), Single Layer Perceptron (SLP), Hopfield Neural Network (HNN).