A Feature Importance Study in Ballet Pose Recognition with OpenPose

Fourie, Margaux; Haar, Dustin van der

doi:10.1007/978-3-030-50334-5_16

Cited by 4 publications

(1 citation statement)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Beberapa penelitian sebelumnya pada dunia pose detection atau penentuan sebuah pose yakni, mendeteksi dan mengklasifikasi isyarat tangan yang menggunakan metode MediaPipe dengan Support Vector Machine (SVM) memiliki akurasi sebesar 99,15 % [8], dan mendeteksi pose balet menggunakan metode OpenPose yang dibantu dengan Support Vector Machine (SVM) untuk klasifikasi. Penelitian ini menunjukkan akurasi dalam pendeteksian dan klasifikasi senilai 99,09% [9] .…”

Section: Pendahuluanunclassified

Implementasi Algoritma Support Vector Machine Terhadap Klasifikasi Pose Balet

Romindo

Barus

Pangaribuan

et al. 2022

bits

View full text Add to dashboard Cite

Ballet is considered as one of the most difficult dance due to its technical posture demanded. If performed without guidance it may cause bad posture to ballerina and some serious injuries. A model in identifying different ballet poses is developed with artificial intelligence in order to tear down this barrier. The main purpose of this paper is to demonstrate a methodology that simplified Ballet Pose Recognition using an opensource framework called MediaPipe and a machine learning algorithm called Support Vector Machine. How the model work is it will pass through two stages: first, it extracts data points from an image dataset using the MediaPipe Pose Estimation library, and then it preprocesses the data, trains, validates, and tests it using the Support Vector Machine algorithm to do some pose classification. The model is trained in seven distinct ballet poses, including First Position, Second Position, Third Position, Fourth Position, Fifth Position, Tendu Devant, and Tendu Derrière. This is purposely done in order to assess the competence of the classification model. An accuracy score of 87% is achieved from the ballet pose classification model and is developed to work on images and live videos.

show abstract

Section: Pendahuluanunclassified