Temporal Reasoning Graph for Activity Recognition
This work addresses activity recognition in videos, which is important for applications like surveillance and human-computer interaction, but it is incremental as it builds on existing graph-based methods for temporal reasoning.
The paper tackles the challenge of recognizing fine-grained and long-term activities in videos by proposing a Temporal Reasoning Graph (TRG) to capture appearance features and temporal relations at multiple scales, achieving state-of-the-art performance on datasets like Something-Something and Charades.
Despite great success has been achieved in activity analysis, it still has many challenges. Most existing work in activity recognition pay more attention to design efficient architecture or video sampling strategy. However, due to the property of fine-grained action and long term structure in video, activity recognition is expected to reason temporal relation between video sequences. In this paper, we propose an efficient temporal reasoning graph (TRG) to simultaneously capture the appearance features and temporal relation between video sequences at multiple time scales. Specifically, we construct learnable temporal relation graphs to explore temporal relation on the multi-scale range. Additionally, to facilitate multi-scale temporal relation extraction, we design a multi-head temporal adjacent matrix to represent multi-kinds of temporal relations. Eventually, a multi-head temporal relation aggregator is proposed to extract the semantic meaning of those features convolving through the graphs. Extensive experiments are performed on widely-used large-scale datasets, such as Something-Something and Charades, and the results show that our model can achieve state-of-the-art performance. Further analysis shows that temporal relation reasoning with our TRG can extract discriminative features for activity recognition.