Audiovisual SlowFast Networks for Video Recognition
This work addresses the challenge of learning unified audiovisual representations for video recognition, which is important for applications like action classification and detection, and it is incremental by building on prior SlowFast networks and neuroscience studies.
The paper tackles the problem of integrated audiovisual perception in video recognition by proposing Audiovisual SlowFast Networks, which fuse audio and visual features at multiple layers and introduce DropPathway for regularization, achieving state-of-the-art results on six video action classification and detection datasets.
We present Audiovisual SlowFast Networks, an architecture for integrated audiovisual perception. AVSlowFast has Slow and Fast visual pathways that are deeply integrated with a Faster Audio pathway to model vision and sound in a unified representation. We fuse audio and visual features at multiple layers, enabling audio to contribute to the formation of hierarchical audiovisual concepts. To overcome training difficulties that arise from different learning dynamics for audio and visual modalities, we introduce DropPathway, which randomly drops the Audio pathway during training as an effective regularization technique. Inspired by prior studies in neuroscience, we perform hierarchical audiovisual synchronization to learn joint audiovisual features. We report state-of-the-art results on six video action classification and detection datasets, perform detailed ablation studies, and show the generalization of AVSlowFast to learn self-supervised audiovisual features. Code will be made available at: https://github.com/facebookresearch/SlowFast.