机译:学习使用深卷积神经网络从新的基于骨架的代表中识别3D人类行动
Cerema Equipe Projet STI 1 Ave Colonel Roche F-31400 Toulouse France|Univ Toulouse UPS IRIT F-31062 Toulouse France;
Cerema Equipe Projet STI 1 Ave Colonel Roche F-31400 Toulouse France;
Univ Toulouse UPS IRIT F-31062 Toulouse France;
Aparnix La Gioconda 4355 10B Santiago Chile;
Univ Carlos III Madrid Appl Artificial Intelligence Res Grp Dept Comp Sci Madrid 28270 Spain|Queen Mary Univ London Sch Elect Engn & Comp Sci London England|Cortex Vis Syst Ltd London England;
object recognition; image colour analysis; image classification; image representation; image motion analysis; learning (artificial intelligence); image sequences; video signal processing; feature extraction; convolutional neural nets; skeleton-based representation; human actions; untrimmed videos; three-dimensional motion representation; 3D action recognition; 3D joint coordinates; human body; skeleton sequences; colour encoding process; skeleton frame; colour-coded representation; complex 3D motions; residual network architecture; image-based representations; 3D motion features; 3D human action recognition; learning model; recognition performance; deep convolutional neural networks; action recognition benchmarks; MSR Action3D; NTU-RGB plus D;
机译:学习使用深度卷积神经网络从新的基于骨骼的表示中识别3D人类动作
机译:3D骨骼运动的时空图像表示,用于深度卷积神经网络的视图不变动作识别?
机译:基于骨架的动作识别和分割的深对齐卷积神经网络
机译:使用时空3D卷积神经网络的基于骨骼的人类动作识别
机译:植物分割通过深度卷积神经网络与转移学习的豆豆植物的植物分割和表型特性
机译:深度卷积神经网络的视图不变动作识别的3D骨骼运动的时空图像表示
机译:学习使用深卷积神经网络从新的基于骨架的代表中识别3D人类行动