Cross-view Action Recognition via Contrastive View-invariant Representation
This addresses the problem of recognizing human actions from unseen viewpoints for applications like surveillance and assisted living, where collecting new training data is impractical.
The paper tackles cross-view action recognition by learning view-invariant features from RGB videos or 3D skeleton data, achieving state-of-the-art performance with accuracies up to 99.9% on datasets like NTU-RGB+D 60.
Cross view action recognition (CVAR) seeks to recognize a human action when observed from a previously unseen viewpoint. This is a challenging problem since the appearance of an action changes significantly with the viewpoint. Applications of CVAR include surveillance and monitoring of assisted living facilities where is not practical or feasible to collect large amounts of training data when adding a new camera. We present a simple yet efficient CVAR framework to learn invariant features from either RGB videos, 3D skeleton data, or both. The proposed approach outperforms the current state-of-the-art achieving similar levels of performance across input modalities: 99.4% (RGB) and 99.9% (3D skeletons), 99.4% (RGB) and 99.9% (3D Skeletons), 97.3% (RGB), and 99.2% (3D skeletons), and 84.4%(RGB) for the N-UCLA, NTU-RGB+D 60, NTU-RGB+D 120, and UWA3DII datasets, respectively.