Back to Search Start Over

Recognizing Actions in Videos from Unseen Viewpoints

Authors :
Piergiovanni, AJ
Ryoo, Michael S.
Source :
CVPR 2021
Publication Year :
2021

Abstract

Standard methods for video recognition use large CNNs designed to capture spatio-temporal data. However, training these models requires a large amount of labeled training data, containing a wide variety of actions, scenes, settings and camera viewpoints. In this paper, we show that current convolutional neural network models are unable to recognize actions from camera viewpoints not present in their training data (i.e., unseen view action recognition). To address this, we develop approaches based on 3D representations and introduce a new geometric convolutional layer that can learn viewpoint invariant representations. Further, we introduce a new, challenging dataset for unseen view recognition and show the approaches ability to learn viewpoint invariant representations.

Details

Database :
arXiv
Journal :
CVPR 2021
Publication Type :
Report
Accession number :
edsarx.2103.16516
Document Type :
Working Paper