Personality Trait Identification Using Unconstrained Cursive and Mood Invariant Handwritten Text

Abstract

Identification of Personality is a complex process. Personality traits are stable over time .Individual's behavior naturally varies from occasion to occasion. But there is a core consistency which defines the true nature. The paper addresses this issue of behavior. Graphology is normally a technique used to identify the traits. Accuracy of this technique depends on how skilled the analyst is. Although human intervention in handwriting analysis has been effective, but it is costly and prone to fatigue. An automation of handwritten text is proposed. Basically we have considered three important features in the direction of orientation of the lines :(i) up hill (ii) down hill (iii) constant line. Edge histogram and bounding boxes was used for feature extraction .Known classifiers like SVM & ANN are used for training and the results were compared. The results were about 98% for SVM & 70% with ANN. The analysis was done using single line.