Fuzzy Clustering Based Car Recognition Algorithm

In this paper, a car recognition algorithm based on fuzzy clustering of multiple features of car image is proposes. Firstly main color model and image pyramid are used to get complete contour of car. Then round rate, Fourier descriptor, direction ratio and circumference ratio are extracted and analyzed in detail. Afterwards all features are combined into a feature vector and fuzzy clustering is used to find the cluster centers of the car classes. Finally similarity between the feature vectors of the input image and the cluster centers of the car classes is used to implement car recognition. The experimental results show that our method has fine accuracy in car recognition.