Papers › Improve Temporal Action Proposals using Hierarchical Context

Improve Temporal Action Proposals using Hierarchical Context

3 Apr 2023Pattern Recognition 2023 4archive 2025-07-28

Qinying Liu, Zilei Wang, Shenghai Rong

Temporal action proposal (TAP) aims to generate accurate candidates of action instances in an untrimmed video. It has been proved that contexts are critically important to this task. In this paper, we propose a novel hierarchical context network (HCN) to further explore the snippet-level and proposal-level contexts, which are used to improve the representations of snippets and proposals, respectively. First, we pinpoint that different scales of snippet-level contexts are not equally important for different action instances. To this end, we incorporate a novel gating mechanism into the U-Net structure to capture the content-adaptive snippet-level contexts. Second, to exploit the proposal-level contexts, we propose a task-specific self-attention model with high efficiency. By stacking multiple attention models, we can deeply explore the proposal-level contexts in a wide range. Finally, to leverage both levels of context, we equip HCN with three branches to evaluate proposals from local to global perspectives. Our experiments on the ActivityNet-1.3 and THUMOS14 datasets show that HCN significantly outperforms previous TAP methods. Additionally, further experiments demonstrate that our method can substantially improve the state-of-the-art action detection performance when combined with existing action classifiers.

PaperPDF

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Action DetectionTemporal Action LocalizationTemporal Action Proposal Generation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Temporal Action Localization ActivityNet-1.3 HCN(I3D features) mAP 35.61 #19 of 33 Archive leaderboard report
Temporal Action Localization ActivityNet-1.3 HCN(I3D features) mAP IOU@0.5 52.51 #19 of 33 Archive leaderboard report
Temporal Action Localization ActivityNet-1.3 HCN(I3D features) mAP IOU@0.75 36.10 #19 of 33 Archive leaderboard report
Temporal Action Localization ActivityNet-1.3 HCN(I3D features) mAP IOU@0.95 7.12 #19 of 33 Archive leaderboard report
Temporal Action Proposal Generation ActivityNet-1.3 HCN AR@100 77.13 #3 of 11 Archive leaderboard report
Temporal Action Proposal Generation ActivityNet-1.3 HCN AUC (val) 68.78 #3 of 11 Archive leaderboard report
Temporal Action Proposal Generation THUMOS' 14 HCH AR@100 50.86 #1 of 3 Archive leaderboard report
Temporal Action Proposal Generation THUMOS' 14 HCH AR@1000 67.34 #1 of 3 Archive leaderboard report
Temporal Action Proposal Generation THUMOS' 14 HCH AR@200 57.56 #1 of 3 Archive leaderboard report
Temporal Action Proposal Generation THUMOS' 14 HCH AR@50 64.28 #1 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Concatenated Skip ConnectionConvolutionMax PoolingReLUU-Net

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections