A Pedestrian Detection System With Weak Classifiers

Tetik Y. E. , BOLAT B.

21st Signal Processing and Communications Applications Conference (SIU), CYPRUS, 24 - 26 April 2013 identifier identifier


In this paper, a pedestrian detection system which uses sliding window approach to detect pedestrians in still digital images is presented. The proposed pedestrian detection system combines weak classifiers in an Adaboost like novel way to create a strong classifier. Besides, rectangle ratios and discrete cosine transform coefficients are used as features with the well-known rectangle differences method.