HyperAI超神经

首页资讯论文教程数据集百科 SOTA LLM 模型天梯 GPU 天梯顶会

中文

HyperAI超神经

Action Classification On Toyota Smarthome

评估指标

CS

CV1

CV2

评测结果

各个模型在此基准测试上的表现结果

模型名称	CS	CV1	CV2	Paper Title	Repository
VPN (RGB + Pose)	60.8	43.8	53.5	VPN: Learning Video-Pose Embedding for Activities of Daily Living
MMNet	70.1	-	-	MMNet: A Model-Based Multimodal Network for Human Action Recognition in RGB-D Videos	-
Dense Trajectories	41.9	20.9	23.7	Improved Dense Trajectory with Cross Streams	-
Separable STA (RGB + Pose)	54.2	35.2	50.3	Toyota Smarthome: Real-World Activities of Daily Living	-
π-ViT	72.9	55.2	64.8	Just Add $\pi$! Pose Induced Video Transformers for Understanding Activities of Daily Living
I3D	53.4	34.9	45.1	Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset
NPL	-	39.6	54.6	Recognizing Actions in Videos from Unseen Viewpoints	-
Trans4SOAR (Pose)	-	-	-	Delving Deep into One-Shot Skeleton-based Action Recognition with Diverse Occlusions
UNIK	64.3	36.1	65.0	UNIK: A Unified Framework for Real-world Skeleton-based Action Recognition
AIRStreams	62.11	-	-	Adaptive Intermediate Representations for Video Understanding	-
I3D + Non Local	53.6	34.3	43.9	Non-local Neural Networks
AssembleNet++	63.6	-	-	AssembleNet++: Assembling Modality Representations via Attention Connections

0 of 12 row(s) selected.