Corpus ID: 232335357

Temporal Context Aggregation Network for Temporal Action Proposal Refinement

@inproceedings{Qing2021TemporalCA,
  title={Temporal Context Aggregation Network for Temporal Action Proposal Refinement},
  author={Zhiwu Qing and Haisheng Su and Weihao Gan and Dongliang Wang and Wei Wu and Xiang Wang and Yu Qiao and Junjie Yan and Changxin Gao and Nong Sang},
  booktitle={CVPR},
  year={2021}
}
Temporal action proposal generation aims to estimate temporal intervals of actions in untrimmed videos, which is a challenging yet important task in the video understanding field. The proposals generated by current methods still suffer from inaccurate temporal boundaries and inferior confidence used for retrieval owing to the lack of efficient temporal modeling and effective boundary context utilization. In this paper, we propose Temporal Context Aggregation Network (TCANet) to generate high… Expand
Transferable Knowledge-Based Multi-Granularity Aggregation Network for Temporal Action Localization: Submission to ActivityNet Challenge 2021
  • Haisheng Su, Peiqin Zhuang, +4 authors Yu Qiao
  • Computer Science
  • ArXiv
  • 2021
TLDR
To train a supervised temporal action localizer, Temporal Context Aggregation Network (TCANet) is adopted to generate high-quality action proposals through “local and global” temporal context aggregation and complementary as well as progressive boundary refinement. Expand
Towards High-Quality Temporal Action Detection with Sparse Proposals
TLDR
The main challenge lies in the large variants of action duration and the ambiguous boundaries for short action instances; nevertheless, quadratic-computational global attention prevents query-based methods to build multi-scale feature maps. Expand
Proposal Relation Network for Temporal Action Detection
TLDR
This technical report presents the solution for temporal action detection task in AcitivityNet Challenge 2021, and builds on BMN, and mainly contains three steps: action classification and feature encoding by Slowfast, CSN and ViViT, and proposal generation. Expand
Weakly-Supervised Temporal Action Localization Through Local-Global Background Modeling
TLDR
This report presents a proposed Local-Global Background Modeling Network (LGBM-Net) that is trained to localize instances by using only video-level labels based on Multi-Instance Learning (MIL) and ensemble multiple models to get the final detection results. Expand
Exploring Stronger Feature for Temporal Action Localization
TLDR
The transformer-based methods can achieve better classification performance than convolution-based, but they cannot generate accuracy action proposals and extracting features with larger frame resolution to reduce the loss of spatial information can also effectively improve the performance of temporal action localization. Expand
Temporal Action Proposal Generation with Transformers
TLDR
An intuitively presented temporal action proposal generation framework with original Transformers, called TAPG Transformer, which consists of a Boundary Transformer and a Proposal Transformer that achieves remarkable performance on the temporal action localization task. Expand
TAda! Temporally-Adaptive Convolutions for Video Understanding
  • Ziyuan Huang, Shiwei Zhang, +4 authors M. Ang
  • Computer Science
  • 2021
TLDR
This work presents Temporally-Adaptive Convolutions (TAdaConv) for video understanding, which shows that adaptive weight calibration along the temporal dimension is an efficient way to facilitate modelling complex temporal dynamics in videos. Expand
Relation Modeling in Spatio-Temporal Action Localization
  • Yutong Feng, Jianwen Jiang, +5 authors Yue Gao
  • Computer Science
  • ArXiv
  • 2021
TLDR
The solution utilizes multiple types of relation modeling methods for spatio-temporal action detection and adopts a training strategy to integrate multiple relation modeling in end-to-end training over the two large-scale video datasets. Expand
RGB Stream Is Enough for Temporal Action Detection
TLDR
It is argued that optical flow is dispensable in high-accuracy temporal action detection and image level data augmentation (ILDA) is the key solution to avoid performance degradation when Optical flow is removed. Expand
A Stronger Baseline for Ego-Centric Action Detection
TLDR
An egocentric video action detection method used in the 2021 EPIC-KITCHENS100 competition hosted in CVPR2021 Workshop is analyzed, showing that classification and proposals are conflict in the same network. Expand
...
1
2
...

References

SHOWING 1-10 OF 46 REFERENCES
BSN: Boundary Sensitive Network for Temporal Action Proposal Generation
TLDR
An effective proposal generation method, named Boundary-Sensitive Network (BSN), which adopts "local to global" fashion and significantly improves the state-of-the-art temporal action detection performance. Expand
Accurate Temporal Action Proposal Generation with Relation-Aware Pyramid Network
TLDR
In RapNet, a novel relation-aware module is introduced to exploit bi-directional long-range relations between local features for context distilling and generates superior accurate proposals over the existing state-of-the-art methods. Expand
Fast Learning of Temporal Action Proposal via Dense Boundary Generator
TLDR
An efficient and unified framework to generate temporal action proposals named Dense Boundary Generator (DBG), which draws inspiration from boundary-sensitive methods and implements boundary classification and action completeness regression for densely distributed proposals. Expand
BMN: Boundary-Matching Network for Temporal Action Proposal Generation
TLDR
This work proposes an effective, efficient and end-to-end proposal generation method, named Boundary-Matching Network (BMN), which generates proposals with precise temporal boundaries as well as reliable confidence scores simultaneously, and can achieve state-of-the-art temporal action detection performance. Expand
Multi-Granularity Generator for Temporal Action Proposal
TLDR
Through temporally adjusting the segment proposals with fine-grained information based on frame actionness, MGG achieves the superior performance over state-of-the-art methods on the public THUMOS-14 and ActivityNet-1.3 datasets. Expand
CTAP: Complementary Temporal Action Proposal Generation
TLDR
This work applies a Proposal-level Actionness Trustworthiness Estimator on the sliding windows proposals to generate the probabilities indicating whether the actions can be correctly detected by actionness scores, and applies CTAP as a proposal generation method in an existing action detector, and shows consistent significant improvements. Expand
TURN TAP: Temporal Unit Regression Network for Temporal Action Proposals
TLDR
A novel Temporal Unit Regression Network (TURN) model, which jointly predicts action proposals and refines the temporal boundaries by temporal coordinate regression, and outperforms state-of-the-art performance on THUMOS-14 and ActivityNet datasets. Expand
Temporal Action Detection with Structured Segment Networks
TLDR
The structured segment network (SSN) is presented, a novel framework which models the temporal structure of each action instance via a structured temporal pyramid and introduces a decomposed discriminative model comprising two classifiers, respectively for classifying actions and determining completeness. Expand
Learning Sparse 2D Temporal Adjacent Networks for Temporal Action Localization
TLDR
The Winner method for HACS Temporal Action Localization Challenge 2019 is introduced, built upon the recent proposed 2D-TAN approach, and is capable of involving more context information for long proposals and further learning discriminative features from them. Expand
Cascaded Pyramid Mining Network for Weakly Supervised Temporal Action Localization
TLDR
This paper designs a two-stage cascaded module with proposed Online Adversarial Erasing (OAE) mechanism, where new and complementary regions are mined through feeding the erased feature maps of discovered regions back to the system. Expand
...
1
2
3
4
5
...