Pytorch implementation of Learning Spatialtemporal Features with 3D Convolutional Networks with temporal depth d=3.
C3D is a Convolution Neural Network with a 3 dimensional (3 x 3 x 3) kernel which learns spatial-temporal information across multiple frames in the same video clip.
- Pytorch 1.0 or above
- OpenCV 4.1 or above
- Python 3.6 or above
- Clone the repo
$ git clone https://github.com/carriex/C3D.git- Install the python dependency packages
$ pip install -r requirements.txt