Study on subtractive clustering video moving object locating method with introduction of eigengap
Zhihai Sun, Wenhui Zhou, Er-tao Li, Hua Zhang · 2012
Considering the problem that the iteration number of video moving object locating method based on subtractive clustering existing a lot of uncertainties, a new video moving object locating method based on subtractive clustering introducing eigengap was proposed. Considering the data density function of subtractive clustering method, affinity matrix and the diagonal matrix was constructed, then eigenvalue and eigengap were both computed using normalized affinity matrix, the iteration number of subtractive clustering video moving object locating method could be automatically determined using the first maximum eigengap value. The extraordinary feature of proposed method was that the affinity matrix and the diagonal matrix could be both determined while computing density value, and then the iteration number could be automatically determined using eigengap. Experiment results showed that the proposed method had a better performance on dealing with the iteration number of subtractive clustering locating method.