Investigation of signature biometric evidence using deep learning for forecasting personality

Abstract

A crucial field of study, personality assessment has applications in marketing, human resources, and psychology. Existing approaches, however, frequently rely on arbitrary self-report surveys, which might introduce biases and constraints. As a result, there is a need for trustworthy and impartial personality testing tools. Using signature biometric data, we suggest a Salp swarm-optimized Siamese neural network (SSO-SNN) in this research to forecast personality attributes. In order to gauge the effectiveness of the suggested strategy, we first collect information from the CIU handwritten database. A median filter is used during pre-processing to reduce noise in the gathered signature photos. The SSO-SNN uses the SSO algorithm's optimization capabilities to speed up the SNN's learning process. Numerous tests are run to determine how effective the suggested strategy is in terms of metrics like accuracy, precision, recall, and f1-score. Experimental outcomes demonstrate the better performance of the suggested SSO-SNN technique in forecasting personality traits when compared to existing approaches.