Comparative analysis of several feature extraction methods in vehicle brand recognition

Abstract

Several feature extraction methods, such as the local energy shape histogram, the local binary pattern model and the gradient histogram, are comparatively used to characterize vehicle face images, and Support Vector Machines (SVM) are proposed to classify vehicle brands. Theoretical analysis and experimental results show that the vehicle brand recognition method based on HOG feature extraction and SVM exceeds the other four methods, and the recognition rate is up to 92.40%.