Recognition of aggressive human behavior based on SURF and SVM

In this paper, we aim to develop a novel decision algorithm of human behavior using both Speeded Up Robust Features (SURF) and PCA techniques. The SURF offers the opportunity to obtain a high level of performance under the constraint of scale variation with low computing coast to form spatio-temporal features. Thus, the PCA algorithm is used to reduce the dimensionality of the provided features to form robust pattern. The latter is performed as an input for training the Support Vector Machine (SVM). This machine is going to be able to classify the aggressive and nonaggressive behaviors. Different tests are conducted on KTH actions datasets. The obtained results have shown that the proposed technique provides more significant accuracy rate in comparison with current techniques as well as it drives more robustness to a dynamic environment.