Dance recognition system using lower body movement

Area of Science:

• Biomechanics
• Computer Science
• Kinesiology

Background:

• Manual film analysis for movement research is time-consuming and difficult.
• Automated identification of complex, full-body movements from film is lacking.
• Pattern recognition and joint location data can enable automated movement description.

Purpose of the Study:

• To develop an automated system for identifying codified dance movements from film.
• To create an algorithm for comparing unknown movements against a database of known dance moves.
• To enable efficient research into movement characteristics and patterns in sports and dance.

Main Methods:

• Utilized lower body kinematic analysis for movement identification.
• Developed a classification and template correlation algorithm.
• Created a database of approximately 90 model dance and Tai Chi Chuan movements.

Main Results:

• Achieved over 97% accuracy in identifying known dance and Tai Chi Chuan movements.
• The system successfully compared unknown moves against the database.
• The program provided kinematic descriptions for both matched and unmatched moves.

Conclusions:

• Automated kinematic analysis offers a highly accurate solution for identifying complex movements in film.
• This technology significantly reduces the time and effort required for movement research.
• The developed system has broad applications in sports, dance, and biomechanical analysis.