Abstract—With the increase in traffic, the demand for driver assistance systems (DAS) has increased. To predict and prevent vehicle accidents, vehicle detection is an essential component of DAS. Vision is the most commonly used human sense while driving and vision sensors (CCD/CMOS) are cheap. Therefore, vision research applied to DAS has been done. In vision research, most vehicle detection methods have database learning (training) as classification algorithms. Well-trained classifiers will have a strong vehicle detection performance. This article presents an improved database learning method for vehicle detection. We use the AdaBoost classifier’s results for feedback input training data. The false-positive results were added to negative training images and true-positive results were added to positive training images. In the experiment results, we proved that our proposed method is better than the existing AdaBoost classifier performance.
Index Terms—Adaboost, vhicle dtection, dtabase taining, har-like feature, fedback.
Jonghwan Kim is from Daegu Gyeongbuk Institute of Science and Technology, Daegu, South of Korea, (e-mail: kimjonhwan@dgist.ac.kr).
Cite: Jonghwan Kim, "Improved Vehicle Detection Method Using Feedback-Ada Boost Learning," International Journal of Computer Theory and Engineering vol. 5, no. 1, pp. 77-80, 2013.
Copyright © 2008-2024. International Association of Computer Science and Information Technology. All rights reserved.
