Personality Detection Based on Tree Drawing Using Convolutional Neural Network

Personality is one of the unique things that humans have. This theory is often used to determine a person's personality condition, including cognitive style. One way to analyze personality is to use psychodiagnostics using the Baum test. This method expresses a person into the image of a tree. This tree image is divided into three large parts, namely the leaf/crown represent the super ego, trunk represents the ego, and the roots represent identity. Currently, the assessment of Baum test results is still carried out manually by psychologists, which takes a long time and is detailed. The Convolutional Neural Network (CNN) method is one of the best methods for classifying image patterns. In this research, we will apply the CNN method to detect personality based on tree image patterns with limitations on the crown/leaf and trunk. Based on the test results, it was found that the CNN method could detect a person's character based on tree image patterns with details of leaf/crown accuracy of 63% and trunk of 79%. The contribution of this research is the use of deep learning to detect a person's personality automatically.