Video segmentation combining depth maps and intensity images

  1. Diego García-Ordás
  2. Óscar García-Olalla Olivera
  3. Laura Fernández Robles
  4. Maite García-Ordás
  5. Enrique Alegre Gutiérrez
Libro:
Actas de las XXXII Jornadas de Automática, Escuela Técnica Superior de Ingeniería, Univesidad de Sevilla: Sevilla, 7 al 9 de septiembre de 2011

Editorial: Universidad de Sevilla

ISBN: 978-84-694-6454-0

Año de publicación: 2011

Páginas: 57

Congreso: Jornadas de Automática (32. 2011. Sevilla)

Tipo: Aportación congreso

Resumen

In this paper several video segmentation techniques for background subtraction were compared and an improvement to one of them have been proposed. The methods evaluated have been ViBe, that uses both intensity images and depth maps, Frame Difference and the Layered Model background subtraction. The image sequences used for compare the segmentation method have been obtained with Kinect and its official libraries OpenNI. The quality of the segmentation have been measured using the scalable discrepancy measures for segmentation evaluation. In our proposal, it is possible to notice that the segmentation obtained is slightly better when the ViBe method is replaced by the Layered method.