To give human-like capabilities to artificial characters, we should equip them with the ability of inferring user states. These artificial characters should understand the users' behaviors through various sensors and respond back using multimodal output. Besides natural multimodal interaction, they should also be able to communicate with multiple users and among each other in multiparty interactions. Previous work on interactive virtual humans and social robots mainly focuses on one-to-one interactions. In this paper, we study tracking and fusion aspects of multiparty interactions. We first give a general overview of our proposed multiparty interaction system and mention how it is different from previous work. Then, we provide the details of the tracking and fusion component including speaker identification, addressee detection and a dynamic user entrance/leave mechanism based on user re-identification using a Kinect sensor. Finally, we present a case study with the system and provide a discussion on the current capabilities, limitations and future work.