Pedestrian Detection with an Improved Adaboost

Tetik Y. E. , BOLAT B.

IEEE International Symposium on INnovations in Intelligent SysTems and Applications (INISTA), Bulgaria, 19 - 21 June 2013 identifier identifier


This paper focuses on improving the performance of Adaboost (Adaptive Boosting) by using weak classifiers that make classification with a confidence score. Single thresholds and nearest neighbor classifiers are used as base classifiers. The proposed method is applied to the problem of pedestrian detection in still images. Haar-like basic features are used to construct weak classifiers.