Strengthening Skeletal Action Recognizers via Leveraging Temporal Patterns
This work addresses the need for more accurate human behavior classification in computer vision, though it is incremental as it builds on existing models.
The paper tackled the problem of improving skeleton-based action recognition by designing two temporal accessories, discrete cosine encoding and chronological loss, which boosted the accuracy of existing models and achieved new state-of-the-art results on two large datasets.
Skeleton sequences are compact and lightweight. Numerous skeleton-based action recognizers have been proposed to classify human behaviors. In this work, we aim to incorporate components that are compatible with existing models and further improve their accuracy. To this end, we design two temporal accessories: discrete cosine encoding (DCE) and chronological loss (CRL). DCE facilitates models to analyze motion patterns from the frequency domain and meanwhile alleviates the influence of signal noise. CRL guides networks to explicitly capture the sequence's chronological order. These two components consistently endow many recently-proposed action recognizers with accuracy boosts, achieving new state-of-the-art (SOTA) accuracy on two large datasets.