UC San Diego

Vehicle classification is currently a widely implemented component in intelligent vehicles, surveillance systems, and traffic monitoring. The major component of vehicle classification is to learn what feature in the images of vehicle provides the most valuable information, which distinguishes different models. In this work, the study of two different famous feature extraction mechanisms and three classifiers is carefully conducted to provide comparison and analysis. The next important component of this work is the investigation of the effect of viewing angle and lighting conditions on the performance of the classifier. The latter is inspired by previous studies on face-recognition systems with different lighting conditions and poses [1].