Home

Awesome

Deep Action Proposals (DAPs) for Videos

Temporal Action Proposals for long untrimmed videos.

DAPs architecture allows to retrieve segments from long videos where it is likely to find actions with high recall very quickly.

Citation

If you find any piece of code valuable for your research please cite this work:

@Inbook{Escorcia2016,
author="Escorcia, Victor and Caba Heilbron, Fabian and Niebles, Juan Carlos and Ghanem, Bernard",
editor="Leibe, Bastian and Matas, Jiri and Sebe, Nicu and Welling, Max",
title="DAPs: Deep Action Proposals for Action Understanding",
bookTitle="Computer Vision -- ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part III",
year="2016",
publisher="Springer International Publishing",
address="Cham",
pages="768--784",
isbn="978-3-319-46487-9",
doi="10.1007/978-3-319-46487-9_47",
url="http://dx.doi.org/10.1007/978-3-319-46487-9_47"
}

If you like this project, give us a :star: in the github banner :wink:.

What is this?

This repo complements our repo with the training pipeline used for our ECCV-2016 work.

I want to use it, Where can I start?

You can find a complete example of our project here. It covers data pre-processing, training and inference.

After training, you can plug-in a model trained here in our clean and simplified inference project.

Installation and Usage

Install

We use conda for deployment. You can create an evironment for this project with the environment-proto-linux-x64.yml.

Dependencies: gcc, CUDA, conda.

Are you a bash user? Take a look at our install.sh script.

Usage

Once all the dependencies are installed, you are ready to go.

Are you bash & modules-env user? Take a look at our activate.sh script.

What is not here?

This repo does not extract the C3D feature vector representation of your videos. However, we try to alleviate your pain installing and assembling all the cumulus of data such that you do not start from scratch.