SST: Single-Stream Temporal Action Proposals

Welcome to the official repo for SST: Single-Stream Temporal Action Proposals!

SST is an efficient model for generating temporal action proposals in untrimmed videos. Analogous to object proposals for images, temporal action proposals provide the temporal bounds in videos where potential actions of interest may lie.

<div class="centered"> <a href="http://vision.stanford.edu/pdf/buch2017cvpr.pdf" target="_blank"_> <img src="https://dl.dropboxusercontent.com/s/pv2mrc0ps09zqu3/sst_modelfig.png" width="590" alt="SST model overview" /> </div> <br/> </a>

Resources

Quick links: [cvpr paper] [poster] [supplementary] [code]

<!-- [[video](https://drive.google.com/file/d/0B_-dKvCH2VL7dGV1ankxWnJVQmM/view?usp=sharing)] -->

Update: if you find this work useful, you may also find our newer work of interest: link to SS-TAD

Please use the following bibtex to cite our work:

@inproceedings{sst_buch_cvpr17,
  author = {Shyamal Buch and Victor Escorcia and Chuanqi Shen and Bernard Ghanem and Juan Carlos Niebles},
  title = {{SST}: Single-Stream Temporal Action Proposals},
  year = {2017},
  booktitle = {CVPR}
  }

As part of this repo, we also include evaluation notebooks, SST proposals for THUMOS'14, and pre-trained model parameters. Please see the code/ and data/ folders for more.

Dependencies

We include a requirements.txt file that lists all the dependencies you need. Once you have created a virtual environment, simply run pip install -r requirements.txt from within the environment to install all the dependencies. Note that the original code was executed using Python 2.7.

<!-- (For Mac OSX users, you may need to run `pip install --ignore-installed numpy six`) -->