Search

Human action recognition models often suffer from achieving both accurate recognition and subject independence when the amount of training data is limited. In this paper, we propose a data-efficient domain adaptation approach to learning a subject-agnostic action recognition classifier. The core component of our approach is a novel data augmentation called Phase Randomization. On the basis of the observation that individual body size is highly correlated with the amplitude component of the motion sequence, we disentangle the individuality and action features by using contrastive self-supervised learning with data augmentation that randomizes only the phase component of the motion sequence. This enables us to estimate the subject label of each motion sequence and to train a subject-agnostic action recognition classifier by performing adversarial learning with the estimated subject labels. We empirically demonstrate the superiority of our method on two different action recognition tasks (skeleton-based action recognition and sensor-based activity recognition). • Data-efficient domain adaptation for subject-agnostic action recognition. • Phase Randomization data augmentation to disentangle action and subject features. • Extensive experiments on two action recognition benchmark tasks. [ABSTRACT FROM AUTHOR]

• Novel deep method using attention model for learnable, task-adaptive time warping. • Pre-training guided by DTW improves discriminative power and stabilizes learning. • Extensive experiments conducted on more than 50 public datasets. • Superior performance demonstrated in both stand-alone and plug-in scenarios. Similarity measures for time series are important problems for time series classification. To handle the nonlinear time distortions, Dynamic Time Warping (DTW) has been widely used. However, DTW is not learnable and suffers from a trade-off between robustness against time distortion and discriminative power. In this paper, we propose a neural network model for task-adaptive time warping. Specifically, we use the attention model, called the bipartite attention model, to develop an explicit time warping mechanism with greater distortion invariance. Unlike other learnable models using DTW for warping, our model predicts all local correspondences between two time series and is trained based on metric learning, which enables it to learn the optimal data-dependent warping for the target task. We also propose to induce pre-training of our model by DTW to improve the discriminative power. Extensive experiments demonstrate the superior effectiveness of our model over DTW and its state-of-the-art performance in online signature verification. [ABSTRACT FROM AUTHOR]