Huntersxsx/Action-Recognition-Learning

动作识别(Action Recognition)常见模型的Pytorch实现

Python

说明

动作识别(Action Recognition)任务中常见的模型Pytorch实现

主要模型

3D卷积类

C3D：Learning Spatiotemporal Features with 3D Convolutional Networks -D.Tran et al, ICCV 2015.
I3D：Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset -J.Carreira et al, CVPR 2017.
P3D：Learning Spatio-Temporal Representation with Pseudo-3D Residual Networks -Z.Qui et al, ICCV 2017.
R(2+1)D：A Closer Look at Spatiotemporal Convolutions for Action Recognition -D.Tran et al， CVPR 2018.
3D ResNets：Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet? -K.Hara et al, CVPR 2019.

Two Stream类

Two Stream：Two-Stream Convolutional Networks for Action Recognition in Videos -K.Simonyan and A.Zisserman, NIPS 2014.
Two Stream Fused：Convolutional Two-Stream Network Fusion for Video Action Recognition -C.Feichtenhofer et al, CVPR 2016.
TSN：Temporal Segment Networks: Towards Good Practices for Deep Action Recognition -L.Wang et al, arXiv 2016

CNN+RNN类

LRCN：Long-term Recurrent Convolutional Networks for Visual Recognition and Description -J.Donahue et al, CVPR 2015.
ConvLSTM：Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting -X Shi et al, NIPS 2015.

参考