The Investigation of Drift and Drift Removal in Markerless Motion Capture Suit in Preparation for Cultural (Aeta) Dance Classification

Abstract

Motion capture is widely used today to capture and analyze human motion. The advent of motion capture systems has given rise to new possibilities in in many applications motion detection and monitoring, including digitization of performing arts activities such as dance. One of the best ways to capture motion is using SmartSuit Pro, a markerless motion capture system. SmartSuit Pro is a wearable suit embedded with inertia sensors in capturing motion. Inertial sensors are considered source less, however, due to a magnetic field’s pervasive presence on earth, this effect makes the magnetic source available almost anywhere. Therefore, this results in the addition of the drift effect to the output model. Drift is one of the main problems found in the horizontal plane of the output motion capture data. Drift is the small error occurring throughout the calculation of angular velocity and acceleration. In the performed Aeta dance recording, drift was encountered because there was an instance that the actor’s foot was not touched the ground where it was supposed to get in contact with it. The drift should be clean as this is an essential step to have an exceptional output of frames in preparation for dance classification. With the use of locomotion and other data filters, identifying and removing data was made possible. The data cleaning was completed and produced a output with no drift.