Sistema Basado en Detección de Pose para el Análisis Computacional de la Ejecución Técnica en Danza Folclórica Mexicana

This study aimed to design and implement an artificial intelligence-based system to analyze technical execution in Mexican folk dance through pose detection and computational comparison techniques. A client–server architecture was developed, consisting of an interaction layer, a computer vision processing layer, and a data storage layer. The system employed MediaPipe Pose technology, utilizing 33 body landmarks to compute joint angles and generate a Movement Similarity Percentage by frame-by-frame comparison with an expert reference model. The research followed a pre-experimental single-group design with pre- and post-intervention measurement. The quantitative phase recorded angular metrics and similarity levels, while the qualitative phase examined participants’ perceptions regarding automated audio feedback. Results demonstrated the technical feasibility of quantifying movement execution without external sensors and highlighted the system’s potential as a pedagogical support tool. It was concluded that integrating computer vision and computational learning allowed objective analysis in a discipline traditionally assessed through subjective observation.