A multimodal biometric system using global features for identical twins identification

Abstract

Pattern recognition studies are currently focusing on twin's biometric identification. The system of twins' biometric Identification can potentially differentiate the individual's biometric pattern. With the new Unimodal biometric identification, identical twins are precisely and reliably identified, with well exposure of certain traits. However, it is much more challenging to identify identical twins as they share so many similarities between them, as opposed to identifying the non-twins. Therefore, this study proposes the use of more than one biometric trait with global features. Pattern recognition application includes extracting and selecting meaningful features and this has brought to the key question in twin handwriting-fingerprint identification: How to obtain features from many writing styles and shapes of twin handwritingfingerprint in order that the reflection of the right person between twins can be obtained. Global with Aspect United Moment Invariant for the extractions of global feature using the identical twin multi-biometric identification is thus proposed by this study.