Analysis of signature for the prediction of personality traits

摘要

The richest information about different emotional states and thinking styles of the person is carried by signature The signature analysis is one of the most effective and reliable indicator for prediction of personality. As it reveals the true personality which includes fears, honesty and many other individual personality traits. This can happen with the help of few features like underscores below signature, appearance of dot on the letter, curved start, ending stroke, and streaks disconnected. This paper is focused on above mentioned features. The database has 60 signatures of 10 different persons with 6 samples of each and is properly scanned at 500dpi resolution scanner. The performance is evaluated by splitting a signature into five categories-left, right, upper, middle and bottom. An artificial neural network and structural identification algorithm is used for the prediction of personality and the obtained accuracy rate is 100%, 95%, 94%, 96% and 92% respectively The significance of work is to reveal personality traits in criminology, medical science and counseling.