Temporal Action Localization On Activitynet

评估指标

mAP

mAP IOU@0.5

mAP IOU@0.75

mAP IOU@0.95

评测结果

各个模型在此基准测试上的表现结果

					Paper Title	Repository
RDFA-S6 (InternVideo2-6B)	42.9	64.1	44.0	10.6	Enhancing Temporal Action Localization: Advanced S6 Modeling with Recurrent Mechanism
ActionMamba (InternVideo2-6B)	42.02	62.43	43.49	10.23	Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding
PRN+BMN (ensemble)	42.0	59.7	-	-	Proposal Relation Network for Temporal Action Detection
AdaTAD (VideoMAEv2-giant)	41.93	61.72	43.35	10.85	End-to-End Temporal Action Detection with 1B Parameters Across 1000 Frames
InternVideo2-6B	41.2	-	-	-	InternVideo2: Scaling Foundation Models for Multimodal Video Understanding
InternVideo2-1B	40.4	-	-	-	InternVideo2: Scaling Foundation Models for Multimodal Video Understanding
UniMD+Sync.	39.83	60.29	-	-	UniMD: Towards Unifying Moment Retrieval and Temporal Action Detection
PRN (CSN)	39.4	57.9	-	-	Proposal Relation Network for Temporal Action Detection
InternVideo	39.00	-	-	-	InternVideo: General Video Foundation Models via Generative and Discriminative Learning
TCANet (SlowFast R101)	37.56	54.33	39.13	8.41	Temporal Context Aggregation Network for Temporal Action Proposal Refinement
PRN (ViViT)	37.5	55.5	-	-	Proposal Relation Network for Temporal Action Detection
AVFusion	36.82	54.34	37.66	8.93	Hear Me Out: Fusional Approaches for Audio Augmented Temporal Action Localization
TriDet (TSP features)	36.8	54.7	38.0	8.4	TriDet: Temporal Action Detection with Relative Boundary Modeling
TadTR (TSP features)	36.75	53.62	37.52	10.56	End-to-end Temporal Action Detection with Transformer
ActionFormer (TSP feautures)	36.6	54.7	37.8	8.4	ActionFormer: Localizing Moments of Actions with Transformers
TAGS (I3D)	36.5	-	-	-	Proposal-Free Temporal Action Detection via Global Segmentation Mask Learning
VSGN (TSP features)	35.94	53.26	36.76	8.12	Video Self-Stitching Graph Network for Temporal Action Localization
TSP	35.81	51.26	37.12	9.29	TSP: Temporally-Sensitive Pretraining of Video Encoders for Localization Tasks
HCN(I3D features)	35.61	52.51	36.10	7.12	Improve Temporal Action Proposals using Hierarchical Context	-
DCAN (TSN features)	35.39	51.78	35.98	9.45	DCAN: Improving Temporal Action Detection via Dual Context Aggregation

0 of 33 row(s) selected.