PEMANFAATAN ESTIMASI POSE GERAK PADA PENARI TRANCE DALAM RITUAL ADAT NGALAKSA

Abstract

This research aims detected the motions poses of trance dancers as a computer vision model in the trance dancers performing arts. The annual ceremonies of the Ngalaksa ritual and the Tarawangsa art as cultural traditions in Rancakalong Tourism Village, located in Sumedang, West Java. This ritual serves as a reverent to the Rice Goddess, with the farmers engaging in ceremonial processions and traditional Ngalaksa rites. Accompanied by the melodic strains of Tarawangsa music, participants dancing in rhythmic dance, some even entering trance states. On the other hand, the advancement of artificial intelligence technology is rapid. This research used machine learning for detecting the motion poses of dancers experiencing trance. The research endeavors to discern motion poses within trance dancers during Ngalaksa rituals for understanding the dynamics motion of traditional Ngalaksa rituals. Machine learning methods applied particularly computer vision utilizing GluonCV model. The study analyzes a dataset comprising motion poses of dancers in Rancakalong Village. Data gathering involves firsthand observation, field participation, and in-depth interviews with informans, including Saehu, dancers and Arak-Arakan participants. The research findings show the accuracy of pose estimation within the Ngalaksa ritual, though refinement of detection models for both trance and non-trance dancers warrants further research.