Moving objects multi-classification based on information fusion

This paper aims to present a new model for multi-classification in video surveillance, based on data fusion. First, features are extracted. Then, a pre-classification is conducted using each feature separately. Second, the obtained posterior probabilities, are combined using the T-conorm operator. At last, the maximum is applied to specify the label of each detected object. The performance of our model was evaluated using two public datasets. In addition, the used number of classes and features were varied, in order to, validate the efficiency of our model. The obtained results showed that our model improved the classification accuracy up to an average of 99% using SVM, also, it outperformed the other methods. (c) 2020 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).