ChoreoSpot: Applying Pose Estimation to Movement Correction in Dance

Abstract

ChoreoSpot, a web app, is a product that explores the practical application of computer visionresearch in pose estimation to performance dance. The app uses a human pose estimation algorithm to help detect errors in dancers’ movements in rehearsal videos, allowing choreographers to more efficiently correct their dancers while preserving the cleanliness and quality of the piece. This paper first takes an entrepreneurial approach in discussing the app’s potential for market success before detailing the process of implementing pose estimation and error detection on a rehearsal video. Quantitative evaluations measuring the accuracy of ChoreoSpot and user feedback indicated that ChoreoSpot reaches a satisfactory performance in error detection, but has the potential to become an indispensable tool in correcting movement in performance dance.