Personality is a unique thing that everyone has. Personalities show how a person acts, both in daily life and at work. Tracking a person's personality is important for the company, many companies conduct tests to get the personality of the prospective employee in accordance with company criteria. Graphology is a technique to analyze one's personality through handwriting. Deep learning is a development from machine learning. One of the most popular Deep Learning methods for image processing is Convolutional Neural Networks (CNN). Combining graphology approach and machine learning can create a computation system that can identify someone’s personality automatically. This research is aimed to identify someone’s personality using graphology approach and deep learning which is CNN. This research is using structural analysis with margin, dot structure, disconnected streak and separated signature features. While for symbol analysis using CNN for curved start, end streak, upper streak, middle streak, shell signature and underline features. The results show that the accuracy of structural analysis approach was up to 78.20-96.20% and the accuracy of symbol analysis with CNN was up to 98.46%. Choosing better configuration and optimization method can improve high accuracy.

Muhammad Reza Aulia

Affiliation

Refine Results

Personality Identification Based on Handwritten Signature Using Convolutional Neural Networks