Action-Recognition-Learning

Action-Recognition-Learning copied to clipboard

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

• R(2+1)D: https://github.com/irhum/R2Plus1D-PyTorch
• 多种模型实现: https://github.com/MRzzm/action-recognition-models-pytorch
• R(2+1)D: https://github.com/kenshohara/3D-ResNets-PyTorch
• C3D: https://github.com/jfzhang95/pytorch-video-recognition
• HOG：https://www.analyticsvidhya.com/blog/2019/09/feature-engineering-images-introduction-hog-feature-descriptor/
• HOF：https://blog.csdn.net/wsp_1138886114/article/details/84400392