Abstract
Recognizing team actions in the behavior of embodied agents has many practical applications and had seen significant progress in recent years. One approach with proven results is based on HMM-based recognition of spatio-temporal patterns in the behavior of the agents. While it had been shown to work on real-world datasets, this approach was found to be brittle.In this paper we present two contributions which together can significantly increase the robustness of teamwork activity recognition. First we introduce a technique to reduce high dimensional continuous input data to a set of discrete features, which capture the essential components of the team actions. Second, we prefix the actual team action recognition with a role recognition module, which allows us to present the recognizer with arbitrarily shuffled input, and still obtain high recognition rates.We validate the improved accuracy and robustness of the team action recognizer on datasets derived from captured real world data.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
