Video Object detection or activity detection or video classification?

Hi guys,

I was wondering if there is any example or at least pull request in progress regarding a PyTorch example with video object detection based on CNN-deep-learning ?
Or maybe similar example for activity detection, classification in videos?

In this way i wanna know how to process/feed the videos to pytorch and do the procedure on them.

Any help is greatly appreciated

Well, I don’t know if there is an example or not. As far as I know video architectures manage audio an video in different tower-streams.
Since video fps are very high they usually pick a sequence of frames (depending on architecture).
But in the dataloader you should be perfectly able to deal with this since you can import opencv to read videos and export output them as torch tensors. There are sequential loaders too (you to read frames in order if your database is already sampled)