3D Convolutional Neural Networks for Human Action Recognition

By Shuiwang Ji, Wei Xu, Ming Yang, Kai Yu. 2010

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Non-local Neural Networks (2017, methodology): “In contrast, feedforward models are achieved by 3D convolutions (C3D) [26, 48] in spacetime, and the 3D filters can be formed by “inflating” [13, 7] pre-trained 2D filters.”
Learning Energy-Based Spatial-Temporal Generative ConvNets for Dynamic Patterns (2019, background): “[39] is a generalization of [22] for dynamic patterns by adopting spatial-temporal ConvNets [3] to capture spatial and temporal features of the video sequences.”
Generating Videos with Scene Dynamics (2016, background): “ortant in video recognition tasks [45], it may be helpful in video generation as well. We explore two different network architectures: One Stream Architecture: We combine spatio-temporal convolutions [14, 40] with fractionally strided convolutions [52, 31] to generate video. Three dimensional convol”
Long-term recurrent convolutional networks for visual recognition and description (2014, background): “A large body of recent work has proposed deep architectures for activity recognition ([16, 33, 13, 2, 1]).”
Value Iteration Networks (2016, methodology): “Recently, CNNs have been applied to reinforcement learning (RL) tasks with visual observations such as Atari games [19], robotic manipulation [16], and imitation learning (IL) [10].”
Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset (2017, background): “Also, they seem to preclude the beneﬁts of ImageNet pre-training, and consequently previous work has deﬁned relatively shallow custom architectures and trained them from scratch [14, 15, 28, 29].”
Two-Stream Convolutional Networks for Action Recognition in Videos (2014, methodology): “The convolutional models of [12, 14] do not decouple spatial and temporal recognition streams, and rely on the motionsensitive convolutional filters, learnt from the data.”
The Kinetics Human Action Video Dataset (2017, methodology): “3D ConvNets [13, 21, 23] seem like a natural approach to video modeling.”
Dense-Captioning Events in Videos (2017, background): “In our paper, F extracts C3D features [15] where δ = 16 frames.”
Unsupervised Learning of Video Representations using LSTMs (2015, background): “This has led to research in 3D convolutional nets (Ji et al., 2013; Tran et al., 2014), different temporal fusion strategies (Karpathy et al., 2014) and exploring different ways of presenting visual information to convolutional nets (Simonyan & Zisserman, 2014a).”
Learning Deconvolution Network for Semantic Segmentation (2015, background): “Convolutional neural networks (CNNs) are widely used in various visual recognition problems such as image classification [17, 24, 25], object detection [8, 10], semantic segmentation [7, 20], visual tracking [13], and action recognition [15, 23].”
Visually Indicated Sounds (2015, background): “We show a sub-sequence of images corresponding to one impact. is closely related to 3D video CNNs [24, 27], as derivatives across channels correspond to temporal derivatives.”
Beyond short snippets: Deep networks for video classification (2015, background): “Encouraged by these positive results several approaches have been proposed recently to apply CNNs to video and action classification tasks [2, 13, 14, 19].”
UvA-DARE (Digital Academic Repository) Self-Supervised Video Representation Learning With Odd-One-Out Networks (positive): “For example, one can use 3D-convolutions [22], recurrent encoders [38], rank-pooling encoders [15] or simply concatenate frames.”
Temporal Segment Networks: Towards Good Practices for Deep Action Recognition (2016, background): “Several works have been trying to design eﬀective ConvNet architectures for action recognition in videos [12,1,13,27,28].”
Financial Time Series Forecasting with Deep Learning : A Systematic Literature Review: 2005-2019 (2019, background): “edontheconvolutionaloperation. Meanwhile,CNNisthemostcommonmodelthatisfrequentlyused forvisionorimageprocessingbasedclassiﬁcationproblems(imageclassiﬁcation,objectdetection, image segmentation, etc.) [62,63,64]. The advantage of the usage ofCNNis thenumberofparameterswhencomparingthevanillaDLmodelss”
Spatiotemporal Residual Networks for Video Action Recognition (2016, background): “In other work, spatiotemporal features are learned by extending 2D ConvNets into time by stacking consecutive video frames [12].”
Learning Spatio-Temporal Features with 3D Residual Networks for Action Recognition (2017, background): “Another approach that captures the spatio-temporal information adopts spatio-temporal 3D convolutional kernels [2, 10, 18] instead of the 2D ones.”
Deep learning in neural networks: An overview (2014, methodology): “An improved version of the method was published later (Ji et al., 2013).”