mlsanityData to 14 Oct 2026

Action-Agnostic Point-Level Supervision for Temporal Action Detection

arXiv 2412.21205 ↗First posted 30 Dec 2024cs.CV cs.AI cs.LGcited by 5

Not on the map. Why?

Abstract (arXiv)

We propose action-agnostic point-level (AAPL) supervision for temporal action detection to achieve accurate action instance detection with a lightly annotated dataset. In the proposed scheme, a small portion of video frames is sampled in an unsupervised manner and presented to human annotators, who then label the frames with action categories. Unlike point-level supervision, which requires annotators to search for every action instance in an untrimmed video, frames to annotate are selected without human intervention in AAPL supervision. We also propose a detection model and learning method to effectively utilize the AAPL labels. Extensive experiments on the variety of datasets (THUMOS '14, FineAction, GTEA, BEOID, and ActivityNet 1.3) demonstrate that the proposed approach is competitive with or outperforms prior methods for video-level and point-level supervision in terms of the trade-off between the annotation cost and detection performance.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 30 of its 49 references that are in the map. First posted on arXiv 30 Dec 2024.

Builds on

30 of 49 references.

References · 30

  1. 2014Going Deeper with Convolutionscited by 4,241
  2. 2017Quo Vadis, Action Recognition? A New Model and the Kinetics Datasetcited by 3,270
  3. 2020Supervised Contrastive Learningcited by 2,829
  4. 2018Real-world Anomaly Detection in Surveillance Videoscited by 577
  5. 2017Soft-NMS -- Improving Object Detection With One Line of Codecited by 382
  6. 2019BMN: Boundary-Matching Network for Temporal Action Proposal Generationcited by 329
  7. 2019G-TAD: Sub-Graph Localization for Temporal Action Detectioncited by 249
  8. 2017Hide-and-Seek: Forcing a Network to be Meticulous for Weakly-supervised Object and Action Localizationcited by 207
  9. 2019The VIA Annotation Software for Images, Audio and Videocited by 167
  10. 2018SoccerNet: A Scalable Dataset for Action Spotting in Soccer Videoscited by 150
  11. 2017UntrimmedNets for Weakly Supervised Action Recognition and Detectioncited by 130
  12. 2017Weakly Supervised Action Localization by Sparse Temporal Pooling Networkcited by 100
18 more
  1. 2018W-TALC: Weakly-supervised Temporal Activity Localization and Classificationcited by 87
  2. 2018AutoLoc: Weakly-supervised Temporal Action Localizationcited by 75
  3. 2019Background Suppression Network for Weakly-supervised Temporal Action Localizationcited by 73
  4. 2020SF-Net: Single-Frame Supervision for Temporal Action Localizationcited by 62
  5. 2021FineAction: A Fine-Grained Video Dataset for Temporal Action Localizationcited by 61
  6. 2019A Context-Aware Loss Function for Action Spotting in Soccer Videoscited by 56
  7. 2021Temporal Action Segmentation from Timestamp Supervisioncited by 48
  8. 2020Adversarial Background-Aware Loss for Weakly-supervised Temporal Activity Localizationcited by 40
  9. 2021Deep Learning-based Action Detection in Untrimmed Videos: A Surveycited by 32
  10. 2022Weakly Supervised Temporal Action Localization via Representative Snippet Knowledge Propagationcited by 32
  11. 2019Action Recognition from Single Timestamp Supervision in Untrimmed Videoscited by 31
  12. 2020D2-Net: Weakly-Supervised Action Localization via Discriminative Embeddings and Denoised Activationscited by 23
  13. 2021Weakly Supervised Action Selection Learning in Videocited by 21
  14. 2015Temporal Localization of Fine-Grained Actions in Videos by Domain Transfer from Web Imagescited by 19
  15. 2021ACM-Net: Action Context Modeling Network for Weakly-Supervised Temporal Action Localizationcited by 19
  16. 2022Weakly-Supervised Temporal Action Detection for Fine-Grained Videos with Hierarchical Atomic Actionscited by 8
  17. 2023Revisiting Foreground and Background Separation in Weakly-supervised Temporal Action Localization: A Clustering-based Approachcited by 8
  18. 2023Weakly-Supervised Action Localization by Hierarchically-structured Latent Attention Modelingcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 4.

5 papers.

  1. 2025ActAlign: Zero-Shot Fine-Grained Video Classification via Language-Guided Sequence Alignmentcited by 3
  2. 2026Boosting Point-supervised Temporal Action Localization via Text Refinement and Alignmentcited by 0
  3. 2026Exploring the Temporal Consistency for Point-Level Weakly-Supervised Temporal Action Localizationcited by 0
  4. 2026Denoise and Align: Diffusion-Driven Foreground Knowledge Prompting for Open-Vocabulary Temporal Action Detectioncited by 0
  5. 2026SoftRerank: Hierarchical Soft Fusion with Candidate-Label Reranking for Long-Tailed Micro-Action Recognitioncited by 0