Preliminary Results on an AdaBoost-Based Strategy for Pattern Recognition in Clouds of Motion Markers
Juan Luis Jiménez-Bascones, Manuel Graña · 2016
This paper deals with the task of point labeling which is involved in optical motion capture tracking systems. We present a brief introduction to the problem and provide a formal definition. At this preliminary stage of the work the main contribution is to tackle the problem as a classification task, using machine learning techniques as AdaBoost to train a set of weak classifiers. First trials over fictitious data are shown, promising interesting results in terms of efficiency and reliability.