Actor Interaction Relation

Actor interaction relation learning focuses on understanding how individuals interact within a group to perform a collective activity, primarily within the context of group activity recognition. Current research emphasizes developing efficient and effective models, exploring architectures like multi-layer perceptrons (MLPs) and transformers to capture both spatial and temporal relationships between actors. These advancements aim to improve the accuracy and robustness of group activity recognition systems, with implications for applications such as video surveillance, human-computer interaction, and sports analytics. The development of novel loss functions, such as contrastive losses, further enhances the ability to distinguish individual actor contributions within group actions.