mlsanityData to 7 Oct 2026

Towards Universal Representation for Unseen Action Recognition

arXiv 1803.08460 ↗First posted 22 Mar 2018cs.CV cs.AI cs.LGcited by 44

Not on the map. Why?

Abstract (arXiv)

Unseen Action Recognition (UAR) aims to recognise novel action categories without training examples. While previous methods focus on inner-dataset seen/unseen splits, this paper proposes a pipeline using a large-scale training source to achieve a Universal Representation (UR) that can generalise to a more realistic Cross-Dataset UAR (CD-UAR) scenario. We first address UAR as a Generalised Multiple-Instance Learning (GMIL) problem and discover 'building-blocks' from the large-scale ActivityNet dataset using distribution kernels. Essential visual and semantic components are preserved in a shared space to achieve the UR that can efficiently generalise to new datasets. Predicted UR exemplars can be improved by a simple semantic adaptation, and then an unseen action can be directly recognised using UR during the test. Without further training, extensive experiments manifest significant improvements over the UCF101 and HMDB51 benchmarks.

Key facts. Cited by 44 papers in the map (data through 7 Oct 2026). First posted on arXiv 22 Mar 2018.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 12; 2022: 13; 2023: 13; 2024: 4; 2025: 2.

44 papers.

Sort:

By area:

  1. 2022Expanding Language-Image Pretrained Models for General Video Recognitioncited by 320
  2. 2022Fine-tuned CLIP Models are Efficient Video Learnerscited by 178
  3. 2023Vita-CLIP: Video and text adaptive CLIP via Multimodal Promptingcited by 90
  4. 2021Elaborative Rehearsal for Zero-shot Action Recognitioncited by 76
  5. 2023VicTR: Video-conditioned Text Representations for Activity Recognitioncited by 28
  6. 2022Audio-visual Generalised Zero-shot Learning with Cross-modal Attention and Languagecited by 27
  7. 2023Zero-shot Skeleton-based Action Recognition via Mutual Information Estimation and Maximizationcited by 24
  8. 2022Multimodal Open-Vocabulary Video Classification via Pre-Trained Vision and Language Modelscited by 22
  9. 2024OmniVid: A Generative Framework for Universal Video Understandingcited by 22
  10. 2023Seeing in Flowing: Adapting CLIP for Action Recognition with Motion Prompts Learningcited by 16
  11. 2021Object Priors for Classifying and Localizing Unseen Actionscited by 15
  12. 2022Cross-modal Representation Learning for Zero-shot Action Recognitioncited by 15
32 more
  1. 2022Deep Unsupervised Key Frame Extraction for Efficient Video Classificationcited by 14
  2. 2023TVTSv2: Learning Out-of-the-box Spatiotemporal Visual Representations at Scalecited by 14
  3. 2023EZ-CLIP: Efficient Zeroshot Video Action Recognitioncited by 14
  4. 2021CLASTER: Clustering with Reinforcement Learning for Zero-Shot Action Recognitioncited by 12
  5. 2023Video Action Recognition with Attentive Semantic Unitscited by 12
  6. 2023Align before Adapt: Leveraging Entity-to-Region Alignments for Generalizable Video Action Recognitioncited by 12
  7. 2022Alignment-Uniformity aware Representation Learning for Zero-shot Video Classificationcited by 11
  8. 2021Transferring Knowledge from Text to Video: Zero-Shot Anticipation for Procedural Actionscited by 10
  9. 2022Continuous Human Action Recognition for Human-Machine Interaction: A Reviewcited by 10
  10. 2023Language-based Action Concept Spaces Improve Video Self-Supervised Learningcited by 10
  11. 2024Zero-Shot Skeleton-based Action Recognition with Dual Visual-Text Alignmentcited by 9
  12. 2021Zero-Shot Action Recognition from Diverse Object-Scene Compositionscited by 8
  13. 2024ActionHub: A Large-scale Action Video Description Dataset for Zero-shot Action Recognitioncited by 8
  14. 2021Discriminative Latent Semantic Graph for Video Captioningcited by 7
  15. 2021Tell me what you see: A zero-shot action recognition method based on natural language descriptionscited by 7
  16. 2023Generating Action-conditioned Prompts for Open-vocabulary Video Action Recognitioncited by 7
  17. 2023Videoprompter: an ensemble of foundational models for zero-shot video understandingcited by 5
  18. 2024Continual Learning Improves Zero-Shot Action Recognitioncited by 5
  19. 2021ZS-SLR: Zero-Shot Sign Language Recognition from RGB-D Videoscited by 4
  20. 2022Universal Prototype Transport for Zero-Shot Action Recognition and Localizationcited by 4
  21. 2023Telling Stories for Common Sense Zero-Shot Action Recognitioncited by 4
  22. 2021Multi-Modal Zero-Shot Sign Language Recognitioncited by 3
  23. 2021PLSM: A Parallelized Liquid State Machine for Unintentional Action Detectioncited by 2
  24. 2021GAN for Vision, KG for Relation: a Two-stage Deep Network for Zero-shot Action Recognitioncited by 2
  25. 2022Global Semantic Descriptors for Zero-Shot Action Recognitioncited by 2
  26. 2022End-to-End Semantic Video Transformer for Zero-Shot Action Recognitioncited by 1
  27. 2022Learning Using Privileged Information for Zero-Shot Action Recognitioncited by 1
  28. 2022REST: REtrieve & Self-Train for generative action recognitioncited by 1
  29. 2021Few-Shot Action Localization without Knowing Boundariescited by 0
  30. 2023Improving Zero-Shot Action Recognition using Human Instruction with Text Descriptioncited by 0
  31. 2025What Can We Learn from Harry Potter? An Exploratory Study of Visual Representation Learning from Atypical Videoscited by 0
  32. 2025GA2-CLIP: Generic Attribute Anchor for Efficient Prompt Tuningin Video-Language Modelscited by 0