Background Suppression Network for Weakly-supervised Temporal Action Localization
This addresses the challenge of localizing actions in videos without frame-level annotations, which is crucial for video analysis applications, but the approach is incremental as it builds on existing weakly-supervised methods by specifically targeting background suppression.
The paper tackles the problem of weakly-supervised temporal action localization, where only video-level labels are available, by proposing a Background Suppression Network (BaS-Net) that introduces an auxiliary background class and uses a two-branch architecture to suppress background frame activations, achieving state-of-the-art performance on THUMOS'14 and ActivityNet benchmarks.
Weakly-supervised temporal action localization is a very challenging problem because frame-wise labels are not given in the training stage while the only hint is video-level labels: whether each video contains action frames of interest. Previous methods aggregate frame-level class scores to produce video-level prediction and learn from video-level action labels. This formulation does not fully model the problem in that background frames are forced to be misclassified as action classes to predict video-level labels accurately. In this paper, we design Background Suppression Network (BaS-Net) which introduces an auxiliary class for background and has a two-branch weight-sharing architecture with an asymmetrical training strategy. This enables BaS-Net to suppress activations from background frames to improve localization performance. Extensive experiments demonstrate the effectiveness of BaS-Net and its superiority over the state-of-the-art methods on the most popular benchmarks - THUMOS'14 and ActivityNet. Our code and the trained model are available at https://github.com/Pilhyeon/BaSNet-pytorch.