Multidisciplinary Approaches to the Archaeological Recognition and Analysis of Ancient Greek Dance Movements

Abstract

Ancient Greek dance, as a significant component of Mediterranean cultural heritage, holds substantial importance for archaeological research. This paper proposes a multidisciplinary approach (integrating archaeology, motion capture technology, and machine learning) for the recognition and analysis of Ancient Greek dance movements. Initially, by simplifying the human skeletal model, a dancer model containing 10 levels of bone nodes is designed, and the precise capture of dance movement data is achieved using a calibration method based on the classic static alignment pose "T-pose." In the motion recognition and analysis phase, adaptive convolution layers are employed to extract dance movement features, combined with a nonlocal mechanism and spatial attention mechanism to accurately identify specific dance movements and their states. Comparative tests demonstrate that this method exhibits high reliability in recognizing corresponding actions of different Ancient Greek dance types, with an F1 value consistently above 0.70. This research not only provides a novel method for the archaeological study of Ancient Greek dance but also offers scientific support for the preservation and transmission of cultural heritage in the Mediterranean region.