Attention-based Temporal Weighted Convolutional Neural Network for Action Recognition
This work addresses the challenge of effective temporal modeling in action recognition, offering an incremental improvement with a novel attention mechanism for video analysis.
The paper tackles the problem of incorporating temporal information into CNNs for human action recognition by proposing an Attention-based Temporal Weighted CNN (ATW), which embeds a visual attention model as temporal weighting to boost performance, achieving substantial gains by focusing on more relevant video segments.
Research in human action recognition has accelerated significantly since the introduction of powerful machine learning tools such as Convolutional Neural Networks (CNNs). However, effective and efficient methods for incorporation of temporal information into CNNs are still being actively explored in the recent literature. Motivated by the popular recurrent attention models in the research area of natural language processing, we propose the Attention-based Temporal Weighted CNN (ATW), which embeds a visual attention model into a temporal weighted multi-stream CNN. This attention model is simply implemented as temporal weighting yet it effectively boosts the recognition performance of video representations. Besides, each stream in the proposed ATW framework is capable of end-to-end training, with both network parameters and temporal weights optimized by stochastic gradient descent (SGD) with backpropagation. Our experiments show that the proposed attention mechanism contributes substantially to the performance gains with the more discriminative snippets by focusing on more relevant video segments.