Designing a neural network for vehicle type classification using extended Haar features

Vehicle type classification method using artificial neural network (ANN) is proposed in this paper. The presented method relies on ANN to recognize between car, motorcycle and non-vehicle images. Extended Haar feature (EHF) was used to extract unique features from motorcycle and car images. These features were used to train four designed ANNs. The focus of this research is to determine the most optimum ANN for future development, which is good in the term of accuracy and speed. The experimental results show that a more complex ANN has a good accuracy but with a slower processing speed. The most optimum ANN gives 94% of accuracy and 51ms of processing speed to classify a single image.