Classification of Emotion Aroused from Painting Using Correspondence Analysis Features of Trained Pictures

Abstract

Nowadays, psychologists have proven that many external factors and psychological damage affect the physical health of individuals. So this requires the attention of all members of a community. Children are more exposed to such injuries and disorders such as autism and ADHD, including those that mainly affect their quality of life. People with these disorders are having trouble handling their surroundings and cannot show their abilities. So over time, people become isolated, unable to flourish their talents. But people who are concerned about the reform and growth of the community are looking for a way to help these children. Among these people are art therapists which try to increase the quality of life for such patients with mentioned disorders by using art techniques. An important work of art is the visual representation of the treatment process. In this way, the person can express emotions, psychological stresses, emotions and stress by exerting an image or a drama. Therefore, we can imagine a refining role for art therapy, which promotes human communication.We have shown that there exist a meaningful relationship between paintings and Labeled images in a way that a meaningful intelligent system could be trained by extracting effective features based on UTA algorithm and selecting important instances using Fast Condensed Nearest Neighbor (FCNN). The results shown that using mentioned approach and Correspondence Analysis (CA) features can improve the classification accuracy to 66.5% in task1 and 64.4% in task2.