Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China.
Sensors (Basel). 2022 Jul 24;22(15):5521. doi: 10.3390/s22155521.
Group activity recognition that infers the activity of a group of people is a challenging task and has received a great deal of interest in recent years. Different from individual action recognition, group activity recognition needs to model not only the visual cues of individuals but also the relationships between them. The existing approaches inferred relations based on the holistic features of the individual. However, parts of the human body, such as the head, hands, legs, and their relationships, are the critical cues in most group activities. In this paper, we establish the part-based graphs from different viewpoints. The intra-actor part graph is designed to model the spatial relations of different parts for an individual, and the inter-actor part graph is proposed to explore part-level relations among actors, in which visual relation and location relation are both considered. Furthermore, a two-branch framework is utilized to capture the static spatial and dynamic temporal representations simultaneously. On the Volleyball Dataset, our approach obtains a classification accuracy of 94.8%, achieving very competitive performance in comparison with the state of the art. As for the Collective Activity Dataset, our approach improves the accuracy by 0.3% compared with the state-of-the-art results.
群体活动识别是指推断一群人的活动，这是一项具有挑战性的任务，近年来受到了广泛关注。与个体动作识别不同，群体活动识别不仅需要建模个体的视觉线索，还需要建模他们之间的关系。现有的方法基于个体的整体特征来推断关系。然而，人体的某些部位，如头部、手部、腿部及其关系，是大多数群体活动中的关键线索。在本文中，我们从不同视角建立基于部分的图。个体内部部分图用于建模个体不同部分的空间关系，而个体间部分图则用于探索参与者之间的基于部分的关系，其中同时考虑了视觉关系和位置关系。此外，我们还利用了一个两分支框架来同时捕获静态空间和动态时间表示。在排球数据集上，我们的方法获得了 94.8%的分类准确率，与现有技术相比具有非常有竞争力的性能。对于集体活动数据集，与现有技术的结果相比，我们的方法将准确率提高了 0.3%。
