2D MULTIPLE MARKER TRACKING FOR MOTION ANALYSIS

• Main Author: Annisa Salsabila Imran Ka, Andi
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/63033
• Institution: Institut Teknologi Bandung

ABSTRACT 2D MULTIPLE MARKER TRACKING FOR MOTION ANALYSIS By Andi Annisa Salsabila Imran Karim NIM: 18316023 (Undergraduate Program in Biomedical Engineering) Motion capture is one of the most commonly used methods of analyzing an athlete's movement while practicing or competing. A motion capture system which consists of commercial cameras, markers, and open-source software is an alternative to the standard motion capture system which tends to be expensive and not portable, making it suitable for the analysis of athletes' movements in the eld . The open source software used to track markers for athletes is generally not capable of continuous tracking when marker cannot be observed visually, so it requires manual correction from the user. Marker's motion information can be used to predict its position when there are no visual information of the marker. This study will analyse a multi-marker tracking algorithm using Kalman Filter and Hungarian Algoritm. Results of this study shows that the proposed algorithm can predict markers position when occludes, but has not succeeded in overcoming the case of identity switch. Keywords: multiple object tracking, occlusion, identity switch, Kalman Filter, Hungarian Algorithm.