DocumentCode :
3519555
Title :
Silhouette extraction based on time-series statistical modeling and k-means clustering
Author :
Hamad, Ahmed Mahmoud ; Tsumura, Norimichi
Author_Institution :
Grad. Sch. of Adv. Integrated Sci., Chiba Univ., Chiba, Japan
fYear :
2011
fDate :
28-28 Nov. 2011
Firstpage :
584
Lastpage :
588
Abstract :
This paper proposes a simple and a robust method to detect and extract the silhouettes from a video sequence of a static camera based on background subtraction technique. The proposed method analyse the pixel history as a time series observations. A robust technique to detect motion based on kernel density estimation is presented. Two consecutive stages of the k-means clustering algorithm are utilized to identify the most reliable background regions and decrease false positives. Pixel and object based updating mechanism is presented to cope with challenges like gradual and sudden illumination changes, ghost appearance, and non-stationary background objects. Experimental results show the efficiency and the robustness of the proposed method to detect and extract silhouettes for outdoor and indoor environments.
Keywords :
feature extraction; image motion analysis; object detection; pattern clustering; statistical analysis; time series; video signal processing; background subtraction technique; ghost appearance; illumination; k-means clustering; kernel density estimation; motion detection; nonstationary background object; object based updating mechanism; pixel based updating mechanism; pixel history; silhouette detection; silhouette extraction; static camera; time-series statistical modeling; video sequence; Adaptation models; Cameras; History; Image color analysis; Kernel; Lighting; Robustness;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pattern Recognition (ACPR), 2011 First Asian Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4577-0122-1
Type :
conf
DOI :
10.1109/ACPR.2011.6166672
Filename :
6166672
Link To Document :
بازگشت