Vehicle logo classification using support vector machine Ensemble
The vehicle logo is a unique mark of vehicle make. It is classification is one of the vehicle identification methods. Vehicle Logos classification is widely used for security in many places such as government building, campus, roads. Most of the existing supervised classification methods are based on Support Vector Machines (SVM), which can yield ideal results. Although SVM can provide good generalization performance, but the classification results of the SVM in particular problem is often far from the theoretically expected level because its implementation in real problem is based on an approximated algorithm. To improve the limited classification performance of the SVM in vehicle logo classification, we propose to use SVM ensemble with bagging. In bagging each individual SVM is trained independently using randomly chosen training samples and then the results of the SVMs are aggregated to make a collective decision. The simulation results show that SVM ensemble outperforms the stand alone SVM method. We used two-dimensional principal component analysis (2DPCA) for logos image feature extraction.
Keywords: Vehicle logo, Classification, Support vector machines ensemble, bagging.
Download Full-Text
ABOUT THE AUTHORS
Wesal Abdallah
Central South University, Changsha, Hunan, P.R.China, 410083
Li Hong
Central South University, Changsha, Hunan, P.R.China, 410083
Wesal Abdallah
Central South University, Changsha, Hunan, P.R.China, 410083
Li Hong
Central South University, Changsha, Hunan, P.R.China, 410083