The system uses biometric signatures for authentication. The dynamic characteristics of the signature were captured using a biometric tablet (such as VERSAT) that recorded features such as x, y coordinates, pen speed, total time to sign, angle of the pen at the time of signing, number of strokes and strokes, and acceleration. When a signature is captured, its important features are extracted, then preprocessed and stored as a template. Templates refer to dynamic feature datasets extracted from signature samples.

After this stage, this feature dataset is preprocessed and used to train an artificial neural network based model, and these trained networks are then stored in a medical database. During the verification phase, the user's signature is checked against the stored model to verify whether the user is real or fake. Also, another added advantage of this approach is that processing (training and validation) is also done on this biometric machine (VERSAT). This saves storage and costs while reducing the carbon footprint, i.e. it is an energy efficient method as it utilizes mobile and handheld biometric devices such as tablets and phones for access.