무용교육을 위한 AI 기반 무용기술 평가시스템 개발 연구 : 도약 동작을 중심으로

초록

This study develops an AI-based dance technique evaluation system for dance education, focusing on two representative leap movements: the Xi-Liao Leg Leap in Chinese classical dance and the Grand Jeté in ballet. The system was designed to integrate video input, pose estimation, key-frame detection, technical indicator calculation, automated scoring, feedback generation, and report export within a unified instructional workflow. Drawing on prior studies in dance pedagogy, motor learning, and AI-based motion analysis, the study proposes a dual-genre evalua- tive framework based on the principle of “shared structure, differentiated content.” The Xi-Liao Leg Leap mode includes eight indicators: split angle, flight time, hip flexion, torso uprightness, landing stability, front knee angle, back knee angle, and minimum preparation knee angle. The Grand Jeté mode includes eight indicators: maximum split angle, flight time, back knee extension, pelvis opening, prepara- tion knee angle, front leg extension, torso stability, and arm line. The system con- verts these indicators into weighted scores and generates interpretable pedagogical feedback and visual reports. Rather than testing a completed intervention, this study presents a development study that articulates system architecture, scoring logic, and educational design. The resulting framework provides a practical foun- dation for future studies on expert agreement, validity, reliability, and classroom application in AI-supported dance education.

키워드