Euclidean Distance Based Offline Signature Recognition System Using Global and Local Wavelet Features

Abstract

Signature recognition is an important requirement of automatic document verification system. Many approaches for signature recognition are found in literature. A novel approach for offline signature recognition system is presented in this paper, which is based on powerful global and local wavelet features (Energy features). The proposed system functions in three stages. Pre-processing stage, which consists of four steps: gray scale conversion, binarization, thinning and fitting boundary box in order to make signatures ready for feature extraction, Feature extraction stage, where totally 59 global and local wavelet based energy features are extracted which are used to distinguish the different signatures. Finally in classification stage, a simple Euclidean distance measure is used as decision tool. The average recognition accuracy obtained using this model ranges from 90% to 100% with the training set of 10 persons to 30 persons.