Awesome

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.

Citation
Requirements
Installation
Usage
Models
Results
[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

Requirements for Python
Requirements for Matlab
Requirements for Caffe and pycaffe (see: Caffe installation instructions)

Installation

git clone ... TODO.

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