@Inbook{Skublewska-Paszkowska:2025,
 author	= {Skublewska-Paszkowska, Maria and Powroznik, Pawel and Lemonari, Marilena and Aristidou, Andreas},
 editor	= {Liarokapis, Fotis and Shehade, Maria and Aristidou, Andreas and Chrysanthou, Yiorgos},
 title		= {Motion Labeling and Recognition: A Case Study on the Zeibekiko Dance},
 bookTitle	= {Interactive Media for Cultural Heritage},
 year		= {2025},
 publisher	= {Springer Nature Switzerland},
 address	= {Cham},
 pages		= {251--266},
 abstract	= {Performing arts is part of our Intangible Cultural Heritage (ICH), a mainspring of humanity's cultural diversity. Folk dance is one of the most important aspects of ICH that is still alive and continues to influence arts in modern days. Digitizing, documenting, and understanding folk dance and other structured movement systems, is important in the larger scheme of cultural forms. Nowadays, modern and sophisticated equipment, such as optical motion capture systems, allow for accurate dance acquisition, advancing our ability to digitally store, curate, study, present, portray, and re-use intangible dance creations. However, motion captured data do not contain labels, annotations, or semantics to assist organization and indexing. In addition, the large diversity of dance motions and their complexity makes automatic motion segmentation and movement recognition challenging, especially for such highly dynamic, heterogeneous, and stylized motions. In this work, we utilize the Spatial Temporal Graph Convolutional Networks (ST-GCN) to recognize particular dance moves, enabling motion segmentation and annotation. ST-GCN takes advantage of the human topology in temporal order, achieving great success in human motion recognition. We use Zeibekiko as our case study; Zeibekiko is a Greek dance, which is included in the Greece's National Inventory of Intangible Cultural Heritage. It is a unique, soulful dance that requires an inner tension, as it is based on the improvisation of the dancer, mainly expressing the performer's emotions (e.g., pride, happiness, sadness), by making their own unique and characteristic steps. The used classification measures indicate that this is an appropriate tool to classify these kinds of dance patterns. Our method can be useful in several applications in the cultural heritage domain, for example in dance motion analysis and indexing to find potential contextual similarities between dances from different countries, as well as e-learning applications, providing detailed step-by-step education on how to imitate specific dance moves (pirouettes), etc.},
 isbn		= {978-3-031-61018-9},
 doi		= {10.1007/978-3-031-61018-9_11},
 url		= {https://doi.org/10.1007/978-3-031-61018-9_11}
}