Euler number based feature extraction technique for gender discrimination from offline Hindi signature using SVM & BPNN classifier

Offline Signature is an exceptional feature which makes any human as unique from other persons, and by their own handwritten signature one person can be identified . Gender identification may be considered as one of the key features for human identification. In this paper, gender discrimination has been proposed by feature extraction method . The proposed framework considers handwritten hindi signature of each individuals as an input for gender detection . Afterwards, several features are extracted from those images. The extracted features and their values are stored as data, which is further classified using Support vector Machine(SVM) & Back Propagation Neural Network (BPNN), seeking to improve performance on the task. The proposed system is divided into two parts. In the first part, several features such as roundness, skewness, kurtosis, mean, standard deviation, area, Euler number, distribution density of black pixel, entropy, equi-diameter, connected component (cc) and perimeter were taken as feature. Then obtained features are divided into two categories. In the first category experimental feature set contains Euler number, whereas in the second category the obtained feature set excludes the same. In this proposed work, exploring a range of architectures, and obtaining a large improvement in state-of-the- art performance on the training dataset, the largest publicly available dataset on the task. In the training dataset, we obtained reports an improvement of 4.7% in gender classification system by the inclusion of Euler number as a feature, instead of usig only BPNN classifier.