Automatic processing of handwritten Persian for just writing letters to Primary school system using artificial neural networks on Android

One of the important issues in the field of pattern recognition, handwriting recognition identifiers that Much research has been done, but in some ways as one of the issues raised. Developing efficient methods for recognizing handwritten ID can the automatic identification letters and numbers listed on the form, and many other uses for the sums Czech helpful the biggest challenge in this context, diversity is a way of drawing ID. One of the first things that neural networks was proposed as an option to solve it, was the name recognition. Today, artificial neural networks are widely used in the recognition and document analysis. Most efforts in the field of recognition of handwritten and printed there are distinct identifiers. Variety of neural network used in this area is significant. Among them are the multi-layer perceptron (MLP), Support Vector Machine (SVM), Self-organizing networks (SOM), associative networks and other types noted. This paper explores the application of neural networks name hand written recognition systems at different stages of first-grade students are assigned. The first part is an introduction to the issues worthy of artificial neural networks and its applications, and then the perceptron artificial neural networks is one of the most important algorithms are introduced. The second part of the application of neural networks in pre-processing input images and the third, the use of neural networks in the classification and the fourth, the applications of neural networks in the classification and the fifth, conclusions and recommendations are presented.

Keywords :artificial neural network, preprocessing, segmentation, classification algorithms KNN