Performance comparison between SVM and K-Means for vehicle counting

Abstract

Support vector machine (SVM) and K-means have been two well known methods used in classification. Choosing an accurate classifier for good features to differentiate between the foreground and background has a significant effect in increasing the accuracy of the detection. This paper presents and analyzes performance comparison between SVM and K-means classifiers for vehicle counting targeted for intelligence transportation systems (ITS) application. In particular, precision and recall have been used for the comparison between the two methods. Five videos from different weather conditions have been used for the testing purposes. SVM shows a better performance in terms of precision and recall.