mlsanityData to 7 Oct 2026

Recognizing Actions in Videos from Unseen Viewpoints

arXiv 2103.16516 ↗First posted 30 Mar 2021cs.CVcited by 18

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. Cited by 18 papers in the map (data through 7 Oct 2026). Builds on 26 of its 46 references that are in the map. First posted on arXiv 30 Mar 2021.

Builds on

26 of 46 references.

References · 26

  1. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  2. 2017Quo Vadis, Action Recognition? A New Model and the Kinetics Datasetcited by 3,259
  3. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  4. 2017The Kinetics Human Action Video Datasetcited by 1,894
  5. 2018SlowFast Networks for Video Recognitioncited by 1,862
  6. 2015Spatial Transformer Networkscited by 1,423
  7. 2018Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognitioncited by 1,199
  8. 2017A Closer Look at Spatiotemporal Convolutions for Action Recognitioncited by 1,139
  9. 2016NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysiscited by 683
  10. 2018Spherical CNNscited by 452
  11. 20183D human pose estimation in video with temporal convolutions and semi-supervised trainingcited by 401
  12. 2019SynSin: End-to-end View Synthesis from a Single Imagecited by 364
14 more
  1. 2018Learning 3D Human Dynamics from Videocited by 315
  2. 2017Learning SO(3) Equivariant Representations with Spherical CNNscited by 244
  3. 2019Camera Distance-aware Top-down Approach for 3D Multi-person Pose Estimation from a Single RGB Imagecited by 131
  4. 2018Charades-Ego: A Large-Scale Dataset of Paired Third and First Person Videoscited by 127
  5. 2017A New Representation of Skeleton Sequences for 3D Action Recognitioncited by 125
  6. 20163D Human Pose Estimation from a Single Image via Distance Matrix Regressioncited by 73
  7. 2015Multi-view 3D Models from Single Images with a Convolutional Networkcited by 57
  8. 2017Skeleton-based Action Recognition with Convolutional Neural Networkscited by 54
  9. 2019Equivariant Multi-View Networkscited by 44
  10. 2014Learning Spatiotemporal Features with 3D Convolutional Networkscited by 43
  11. 2015Maximum-Margin Structured Learning with Deep Networks for 3D Human Pose Estimationcited by 37
  12. 2018Fine-grained Activity Recognition in Baseball Videoscited by 20
  13. 2018Geometry-Aware Recurrent Neural Networks for Active Visual Recognitioncited by 20
  14. 2016Invariant recognition drives neural representations of action sequencescited by 2

Built on by

Cited by year

Citing papers per year: 2021: 3; 2022: 4; 2023: 4; 2024: 3; 2025: 4.

18 papers.

Sort:

  1. 2023Learning Fine-grained View-Invariant Representations from Unpaired Ego-Exo Videos via Temporal Alignmentcited by 46
  2. 2021UNIK: A Unified Framework for Real-world Skeleton-based Action Recognitioncited by 35
  3. 2022ViA: View-invariant Skeleton Action Representation Learning via Motion Retargetingcited by 22
  4. 2021ViewCLR: Learning Self-supervised Video Representation for Unseen Viewpointscited by 20
  5. 2023Just Add π! Pose Induced Video Transformers for Understanding Activities of Daily Livingcited by 19
  6. 2023DVANet: Disentangling View and Action Features for Multi-View Action Recognitioncited by 19
  7. 2022Learning Viewpoint-Agnostic Visual Representations by Recovering Tokens in 3D Spacecited by 13
  8. 2022Weakly-Supervised Online Action Segmentation in Multi-View Instructional Videoscited by 12
  9. 2023Self-Supervised Video Representation Learning via Latent Time Navigationcited by 11
  10. 2024POV: Prompt-Oriented View-Agnostic Learning for Egocentric Hand-Object Interaction in the Multi-View Worldcited by 10
  11. 2024MV2MAE: Multi-View Video Masked Autoencoderscited by 6
  12. 2022Transformers in Action Recognition: A Review on Temporal Modelingcited by 3
6 more
  1. 2024KAN-HyperpointNet for Point Cloud Sequence-Based 3D Human Action Recognitioncited by 3
  2. 2025Feature Hallucination for Self-supervised Action Recognitioncited by 3
  3. 2021Cross-modal Manifold Cutmix for Self-supervised Video Representation Learningcited by 0
  4. 2025Towards Generalizing Temporal Action Segmentation to Unseen Viewscited by 0
  5. 2025ViewBridge: Curriculum Knowledge Distillation for Activity View-Invariance Under Extreme Viewpoint Changescited by 0
  6. 2025Probabilistic Temporal Masked Attention for Cross-view Online Action Detectioncited by 0