Effectiveness Investigation of Drawing Exactness Assessment Method in Portrait Drawing Learning Assistant System

Portrait drawing is important for developing drawing skills and humanistic sensibilities. It is challenging for novices to draw portraits because it is difficult to master facial proportions and structures without professional guidance. For this reason, we developed a Portrait Drawing Learning Assist System (PDLAS) to help novices by providing auxiliary lines for facial features using OpenPose and OpenCV. Novices can use these auxiliary lines to draw portraits on the iPad. To improve the feedback, we implemented an evaluation method using the local Normalized Cross-Correlation (NCC) algorithm, which calculates similarity scores between the original face photo and the drawing image. It focuses on specific regions such as the eyes, nose, mouth, eyebrows, and hair. In this paper, we asked 3 students at Okayama University in Japan to use PDLAS to draw portraits for evaluation. We prompted users to refine low-scoring facial features and reevaluate to observe improvements in their overall scores.