Identification of Feature Values and Proposal of an Evaluation Model for Assessing Skillfulness of Illustrations Composed of Geometric Figures
Abstract: The purpose of this research, which aims to support skill improvement of beginners in illustration, is to identify feature values and propose a rating scale to evaluate the skillfulness of an illustration that a person has drawn by copying an illustration consisting of geometric figures. Copying is a basic form of practice for illustration. However, it is difficult for learners themselves to evaluate objectively the illustrations they have drawn. In this research, in order to solve this problem, we propose a rating scale for skillfulness of illustrations that reflects the evaluation criteria of experienced illustrators. We interviewed experienced illustrators about their methods of evaluating illustrations, and identified the feature values that relate to the skillfulness of illustrations. These feature values were used as explanatory variables, and multiple regression analysis was applied to construct a rating scale that can express the skillfulness of illustrations by using the subjectively evaluated score of skillfulness as the objective function. When cross-validation was applied to the proposed model, the results generated from the model were closer to the results generated from the rating scale of experienced illustrators than to those generated from the rating scale of beginners. Therefore, it is evident that our model is more accurate than evaluation conducted by beginners.