Authentication System Based on Hand Writing Recognition

In this paper, a proposed method to authentication system based on a hand writing recognition without segmentation to sub letters based on feature extraction Speed Up Robust Features (SURF) and Support Vector Machine (SVM) to enhance the accuracy. After feature extraction, the result can be clustered by using the K-means algorithm to standardize the number of features. These two operations together (feature extraction and feature clustering) called Bag of Word (BOW). It converts the arbitrary number of image features to uniform length features vector. The proposed method experimented using two databases; the standard (IFN/ENIT) database and the proposed database consist of 30 different writers with 70 margins images. The proposed system achieves (96.67%) recognition rate for IFN/IENT database.