Papers › TadML: A fast temporal action detection with Mechanics-MLP

TadML: A fast temporal action detection with Mechanics-MLP

7 Jun 2022arXiv:2206.02997archive 2025-07-28

Bowen Deng, Dongchang Liu

Temporal Action Detection(TAD) is a crucial but challenging task in video understanding.It is aimed at detecting both the type and start-end frame for each action instance in a long, untrimmed video.Most current models adopt both RGB and Optical-Flow streams for the TAD task. Thus, original RGB frames must be converted manually into Optical-Flow frames with additional computation and time cost, which is an obstacle to achieve real-time processing. At present, many models adopt two-stage strategies, which would slow the inference speed down and complicatedly tuning on proposals generating.By comparison, we propose a one-stage anchor-free temporal localization method with RGB stream only, in which a novel Newtonian Mechanics-MLP architecture is established. It has comparable accuracy with all existing state-of-the-art models, while surpasses the inference speed of these methods by a large margin. The typical inference speed in this paper is astounding 4.44 video per second on THUMOS14. In applications, because there is no need to convert optical flow, the inference speed will be faster.It also proves that MLP has great potential in downstream tasks such as TAD. The source code is available at https://github.com/BonedDeng/TadML

PaperPDFCode

Code

boneddeng/tadml officialmentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

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

Tasks

Action DetectionOptical Flow EstimationTemporal Action LocalizationTemporal Localization

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Action Detection THUMOS' 14 TadML-two stream mAP 59.7 #2 of 4 Archive leaderboard report
Action Detection THUMOS' 14 TadML-rgb mAP 53.46 #4 of 4 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(two-stream) Avg mAP (0.3:0.7) 59.70 #14 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(two-stream) mAP IOU@0.3 73.29 #14 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(two-stream) mAP IOU@0.4 69.73 #14 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(two-stream) mAP IOU@0.5 62.53 #14 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(two-stream) mAP IOU@0.6 53.36 #14 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(two-stream) mAP IOU@0.7 39.60 #14 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(rgb-only) Avg mAP (0.3:0.7) 53.46 #20 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(rgb-only) mAP IOU@0.3 68.78 #20 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(rgb-only) mAP IOU@0.4 64.66 #20 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(rgb-only) mAP IOU@0.5 56.61 #20 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(rgb-only) mAP IOU@0.6 45.40 #20 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 TadML(rgb-only) mAP IOU@0.7 31.88 #20 of 42 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

Average PoolingDense ConnectionsDropoutGlobal Average PoolingLayer NormalizationMLP-MixerResidual ConnectionSPEED

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