Hi, I am interested in that how to use the existed MatConvnet to realize 3D-CNN to learn spatio-temporal Features from Video sequence, as described in following papers? the key step is how to realize 3D-Convoution and 3D-pooling.
3D convolutional neural networks for human action recognition-2013-ICMI;
Learning Spatiotemporal Features with 3D Convolutional Networks_2015-ICCV
Could you please give an answer?
Hi, I am interested in that how to use the existed MatConvnet to realize 3D-CNN to learn spatio-temporal Features from Video sequence, as described in following papers? the key step is how to realize 3D-Convoution and 3D-pooling. 3D convolutional neural networks for human action recognition-2013-ICMI; Learning Spatiotemporal Features with 3D Convolutional Networks_2015-ICCV Could you please give an answer?