Haralick Feature Descriptors for Gender Classification Using Fingerprints: A Machine Learning Approach

This Gender classification is a key task for human beings, as many social interactions are gender-based. The research carried out so for on gender classification techniques can be vigorously estimated from the face image with relatively more complex resulting in high accuracy. Fingerprint biometric trait is another technique which is used for the estimation of gender information and classification. It is a significant step in anthropology in order to identify the femininity of a criminal and shorten the list of suspects search. In this work, Haralick texture features are used to extract the gender information from fingerprints for classification of male and female. The experiment is conducted on fingerprints collected from different age groups of rural and urban people. According to the experimental observations a 92% and 94% classification rate is achieved for linear discriminant analysis (LDA) and quadratic discriminant analysis (QDA) classifiers respectively. The work has been analyzed and the results reported in this are found to be satisfactory and more competitive. General Terms- Image Processing, Pattern Recognition.