Title :
A pedestrian detection method based on Dirichlet distribution and histogram of components
Author :
Xu Li ; Xiaoru Wang ; Jia Long ; Lingpu Gong
Author_Institution :
Beijing Key Lab. of Intell. Telecommun. Software & Multimedia, Beijing Univ. of Posts & Telecommun., Beijing, China
fDate :
Oct. 30 2012-Nov. 1 2012
Abstract :
Pedestrian detection is a hot topic in computer vision recently. It is a challenging issue to detect the pedestrian in images accurately and quickly. In this paper, we propose a quick and accurate pedestrian detection method based on Dirichlet distribution and histogram of components. This method can find out the pedestrian in images accurately. Because of the introduction of Dirichlet distribution and histogram of components, our method reduces the time cost on detecting pedestrian. The experiment also showed that it can detect the pedestrian sheltered from other object or in the complicated background.
Keywords :
computer vision; object detection; statistical analysis; Dirichlet distribution; computer vision; histogram-of-components; pedestrian detection method; time cost reduction; Detectors; Dictionaries; Feature extraction; Head; Histograms; Support vector machines; Training; Bag-of-words; Component; Dirichlet distribution; Pedestrian detection; SVM classifier; Sift;
Conference_Titel :
Cloud Computing and Intelligent Systems (CCIS), 2012 IEEE 2nd International Conference on
Conference_Location :
Hangzhou
Print_ISBN :
978-1-4673-1855-6
DOI :
10.1109/CCIS.2012.6664318
