Video dataset of Balinese dance basic movement for action recognition

Area of Science:

• * Digital Humanities
• * Cultural Heritage Studies
• * Computer Vision

Background:

• * Traditional Balinese dance is a significant intangible cultural heritage.
• * Existing documentation methods may not capture the full complexity of dance movements.
• * Computational approaches offer new avenues for preserving and analyzing cultural practices.

Purpose of the Study:

• * To present a detailed, multi-angle video dataset of traditional Balinese dancing.
• * To facilitate computational analysis and preservation of this cultural art form.
• * To highlight gender-based variations in dance movements.

Main Methods:

• * Collected 1740 (female) and 828 (male) high-definition video recordings of 24 distinct Balinese dance movements.
• * Utilized multiple smartphone cameras from various perspectives with three strategically placed cameras.
• * Performed systematic preprocessing, including segmentation and labeling, for clarity and research utility.

Main Results:

• * A comprehensive dataset featuring 24 traditional Balinese dance movements.
• * Videos categorized by dancer gender, dance type, and camera angle.
• * Emphasis on gender-specific movement variations for in-depth cultural analysis.

Conclusions:

• * The dataset serves as a valuable resource for cultural studies and computational research.
• * This initiative bridges technological advancements with the preservation of intangible cultural heritage.
• * The structured data supports detailed analysis of traditional dance forms.