Skip to content
Code and models of paper " Chained Multi-stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection" in ICCV 2017
Branch: master
Clone or download
Latest commit 108a2f4 Aug 10, 2018
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
body_part_segmentation Update README.md Aug 10, 2018
files Add files via upload May 3, 2018
scripts Add files via upload May 3, 2018
LICENSE Initial commit Jan 26, 2018
README.md Update README.md May 3, 2018

README.md

Code and Models for "Chained Multi-stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection" in ICCV 2017.

By Mohammadreza Zolfaghari, Gabriel L. Oliveira, Nima Sedaghat, Thomas Brox

Update

  • 2018.4.30: Scripts for creating body-part mask to train body-part segmentation network.
  • 2018.2.26: The pretrained models and scripts for creating human pose maps are released.

Contents

  1. Citation
  2. Requirements
  3. Installation
  4. Usage
  5. Models
  6. Results
  7. [Project page](#Project page)

Citation

If you find ChainedNet useful in your research, please consider to cite:

    @InProceedings{ZOSB17a,
    author       = "Mohammadreza Zolfaghari and
                Gabriel L. Oliveira and
                Nima Sedaghat and
                Thomas Brox",
    title        = "Chained Multi-stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection",
    booktitle    = "IEEE International Conference on Computer Vision (ICCV)",
    month        = " ",
    year         = "2017",
    url          = "http://lmb.informatik.uni-freiburg.de/Publications/2017/ZOSB17a"
    }

Requirements

  1. Requirements for Python
  2. Requirements for Matlab
  3. Requirements for Caffe and pycaffe (see: Caffe installation instructions)

Installation

  1. git clone ... TODO.

  2. Build Caffe and pycaffe

    cd $caffe_FAST_ROOT/
    # Now follow the Caffe installation instructions here:
    # http://caffe.berkeleyvision.org/installation.html
    make all -j8 && make pycaffe && make matcaffe

Usage

After successfully completing the installation, you are ready to run all the following experiments.

Part 0: Network Inputs

  • RGB, use extract_frames_frmRate.sh in the scripts folder to extract frames.

  • Optical Flow: TVL1, Brox

  • Pose

         Inputs            
    

Part 1: Body Part Segmentation

Please follow steps explained in Body Part Segmentation

Image+mask BodyPart mask

Part 2: Training the Chained Multi-stream network

Note: TODO

Part 3: Results

Note: TODO

Recognition Detection

Project page

https://lmb.informatik.uni-freiburg.de/projects/action_chain/

Contact

Mohammadreza Zolfaghari

Questions can also be left as issues in the repository. We will be happy to answer them.

You can’t perform that action at this time.