A study of neural network and multi-channel handwritten isolated digit recognition.

Abstract

Recognition of handwritten digits/characters is an important and necessary step in many documents processing application and this has been a popular topic of research for many years. Many approaches and significant research efforts have been proposed and made to solve this problem. Some use only the pixel-image as input to a powerful statistical or neural classifier. Others preprocess the data in order to extract some features that are fed into a classifier. The main difficulty lie in the handwritten recognition task is to tell which digit is represented by a pixel-image of a handwritten digits/character, this is due to the high variability of the scanned image caused by the particular writing style of different persons and also the scanning quality.