Cross-Domain First Person Audio-Visual Action Recognition through Relative Norm Alignment
This work addresses cross-domain generalization issues in first person action recognition for applications with wearable cameras, but it is incremental as it builds on existing multi-modal architectures.
The paper tackles the problem of cross-domain first person action recognition by addressing environmental bias in learned representations, proposing an audio-visual loss that aligns feature norms to improve generalization, and demonstrates strong results on the EPIC-Kitchens dataset.
First person action recognition is an increasingly researched topic because of the growing popularity of wearable cameras. This is bringing to light cross-domain issues that are yet to be addressed in this context. Indeed, the information extracted from learned representations suffers from an intrinsic environmental bias. This strongly affects the ability to generalize to unseen scenarios, limiting the application of current methods in real settings where trimmed labeled data are not available during training. In this work, we propose to leverage over the intrinsic complementary nature of audio-visual signals to learn a representation that works well on data seen during training, while being able to generalize across different domains. To this end, we introduce an audio-visual loss that aligns the contributions from the two modalities by acting on the magnitude of their feature norm representations. This new loss, plugged into a minimal multi-modal action recognition architecture, leads to strong results in cross-domain first person action recognition, as demonstrated by extensive experiments on the popular EPIC-Kitchens dataset.