C3D Model for Keras trained over Sports 1M

Author: Alberto Montes

Advisors: Xavier Giro and Amaia Salvador

We provide an adaptation to Keras of the C3D model used with a fork of Caffe, which was trained over the Sports1M dataset.

Details about the network architecture can be found in the following paper:

Tran, Du, Lubomir Bourdev, Rob Fergus, Lorenzo Torresani, and Manohar Paluri. "Learning Spatiotemporal Features With 3D Convolutional Networks." In Proceedings of the IEEE International Conference on Computer Vision, pp. 4489-4497. 2015.

 Find our model and an example of how to use on this page on Github Gist.