A B S T R A C TIn this study, a strategy for feature fusion for team behaviors recognition using automatically generated RBF neural network is proposed, for various features need to be extracted in the course of team behaviors recognition and it is difficult to estimate the contribution of various features for identifying and decrypting team behaviors. The burden of high-level recognition algorithm is use eased by using the underlying features of moving target, such as the trajectory characteristics extracted by using the method of trajectory growth. KPCA algorithm is used to nonlinearly reduce dimensionality of extracted characteristics before feature fusion for extracted characteristics of team behaviors. Automatically generated RBF neural network is constructed and feature fusion is realized by using Dempster-Shafer combination rules and network learning. Parameters µ, α and γ are obtained through networks learning in the course of features fusion, s and p is decided by the decreased gradient of output error. The accuracy of behavior recognition is increased dramatically and the processing time is shorten significantly.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.