An Extension of Drawing Exactness Assessment Method to Hair Evaluation in Portrait Drawing Learning Assistant System

Portrait drawing learning is important for developing painting skills and humanistic sensibilities to a lot of people. However, it is challenging for novices to start learning portrait drawing, because it is often difficult to master facial proportions and structures without professional guidance. To solve this issue, we have developed a Portrait Drawing Learning Assist System (PDLAS) to help novices draw portraits by providing auxiliary lines for facial features that are generated using OpenPose and OpenCV from the face photo. Besides, we have implemented the exactness assessment method to evaluate drawing accuracy using the Local Normalized Cross-Correlation (NCC) algorithm. It calculates the similarity score between the original face photo and the drawing result. However, the current one is limited to evaluate the eyes, nose, mouth, and eyebrows in a face. In this paper, we present an extension of the drawing exactness assessment method to cover the hair for comprehensive feedback to users in PDLAS. For evaluations, we applied the proposal to drawing results by six students at Okayama University, Japan, using PDLAS and confirmed the validity.