… convolution into a 2D spatial convolution followed by 1D temporalconvolution, which we name (2+1)D convolution… chitecture [13], (2+1)D convolutions lead to state-of-theart results on 4 …
Y Zhao, Y Xiong, D Lin - Advances in neural information …, 2018 - proceedings.neurips.cc
… into separate components respectively for spatial and temporalconvolutions. The … in the actionrecognition literature, we introduce the concept of trajectory convolution. In trajectory …
K Liu, W Liu, C Gan, M Tan, H Ma - … of the AAAI conference on artificial …, 2018 - ojs.aaai.org
… Video-based actionrecognition with deep neural networks has shown … actionrecognition architecture, called Temporal Convolutional 3D Network (T-C3D), which learns video action …
… To learn spatial and temporalconvolution kernels effectively, we follow the ideas from [28] and … In Table 1, we also compare with [10] in the setting of using temporalconvolution pipeline …
A Stergiou, R Poppe - 2019 18th IEEE International Conference …, 2019 - ieeexplore.ieee.org
… with a 3D convolution. Therefore, we argue that spatio-temporalconvolutions of video data with … for the current availability of training data per action class. We observe that the motion in …
… of temporal models, which we call Temporal Convolutional Networks (TCNs), that use a hierarchy of temporalconvolutions to perform fine-grained action … long-range temporal patterns …
L Jing, X Yang, Y Tian - Journal of Visual Communication and Image …, 2018 - Elsevier
… , to capture the overall temporal dynamics from an entire video in a single process … action recognition. It remains an open question for actionrecognition on how to deal with the temporal …
T Lin, X Zhao, Z Shou - arXiv preprint arXiv:1707.06750, 2017 - arxiv.org
… in both temporalaction proposal and temporalaction localization task. Our results suggested that anchor mechanisms and temporalconvolution can work well in temporalaction …
… spatialtemporal structure, we apply temporalconvolution on the local spatial-temporal … employs a separate channel-wise and temporal-wise convolution over the feature sequence of a …