The Effects of Using a Noise Filter and Feature Selection in Action Recognition: An Empirical Study
Carolina Gabriela Maldonado Mendez, Sergio Hernández Méndez, Ana Luisa Solís, Homero Vladimir Ríos-Figueroa, Antonio Marı́n-Hernández · 2017
In this paper we are interested in knowing, if a filter and a feature selection are used, the action recognition is improved. In the filter phase, Kalman filter and Savitzky-Golay filter are proposed to analyse skeleton acquired by the Kinect, in order to minimize the noise. In the phase of feature selection, the algorithm Reduction of Feature Dimensions based on Standard Deviation (RFD-SD) and Principal Component Analysis(PCA) are proposed to select the relevant set of features. The obtained results suggest that action recognition improves by using Kalman filter and the algorithm RFD-SD.