Multi Person Pose Estimation
Matlab Other
Switch branches/tags
Nothing to show
Clone or download
Eldar Insafutdinov
Eldar Insafutdinov Updated submodule of the solver
Latest commit 096e2d1 Aug 30, 2016
Permalink
Failed to load latest commit information.
data Initial commit Jun 24, 2016
external Updated submodule of the solver Aug 30, 2016
lib updated unix cmd Aug 3, 2016
models Initial commit Jun 24, 2016
.gitignore Initial commit Jun 24, 2016
.gitmodules Initial commit Jun 24, 2016
README.md updated readme Jul 20, 2016
start_matlab.sh Initial commit Jun 24, 2016
start_matlab_console.sh Initial commit Jun 24, 2016
startup.m Initial commit Jun 24, 2016

README.md

Deep(er)Cut: Multi Person Pose Estimation

This short documentation describes steps necessary to compile and run the code that implements DeepCut and DeeperCut papers:

Leonid Pishchulin, Eldar Insafutdinov, Siyu Tang, Bjoern Andres, Mykhaylo Andriluka, Peter Gehler, and Bernt Schiele
DeepCut: Joint Subset Partition and Labeling for Multi Person Pose Estimation
In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016

Eldar Insafutdinov, Leonid Pishchulin, Bjoern Andres, Mykhaylo Andriluka, and Bernt Schiele
DeeperCut: A Deeper, Stronger, and Faster Multi-Person Pose Estimation Model
In European Conference on Computer Vision (ECCV), 2016
For more information visit http://pose.mpi-inf.mpg.de

Prerequisites

Installation Instructions

  1. Clone repository

    $ git clone https://github.com/eldar/deepcut --recursive
    
  2. Build Caffe and its MATLAB interface after configuring Makefile.config

    $ cd external/caffe
    $ make -j 4 all matcaffe
    
  3. Build liblinear, specify the path to the MATLAB installation

    $ cd external/liblinear-1.94/matlab
    $ CC=gcc CXX=g++ MATLABDIR=/usr/lib/matlab-8.6/ make
    
  4. Build solver

    $ cd external/solver
    $ cmake . -DGUROBI_ROOT_DIR=/path/to/gurobi603/linux64 -DGUROBI_VERSION=60
    $ make solver-callback
    
  5. Obtain Gurobi license from http://www.gurobi.com/downloads/licenses/license-center and place the license file license.lic in data/gurobi or modify parameter p.gurobi_license_file in lib/pose/exp_params.m to point to the license file location

Download models

$ cd data
$ ./download_models.sh

Run Demo

$ cd <root_dir>
$ ./start_matlab.sh
% in MATLAB
>> demo_multiperson

CNN-based part detectors

Access DeeperCut Part Detectors to download stand-alone part detectors that produce dense scoremaps.

Citing

@inproceedings{insafutdinov2016deepercut,
	author = {Eldar Insafutdinov and Leonid Pishchulin and Bjoern Andres and Mykhaylo Andriluka and Bernt Schieke},
	title = {DeeperCut: A Deeper, Stronger, and Faster Multi-Person Pose Estimation Model},
	booktitle = {European Conference on Computer Vision (ECCV)},
	year = {2016},
	url = {http://arxiv.org/abs/1605.03170}
    }
@inproceedings{pishchulin16cvpr,
	author = {Leonid Pishchulin and Eldar Insafutdinov and Siyu Tang and Bjoern Andres and Mykhaylo Andriluka and Peter Gehler and Bernt Schiele},
	title = {DeepCut: Joint Subset Partition and Labeling for Multi Person Pose Estimation},
	booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
	year = {2016},
	url = {http://arxiv.org/abs/1511.06645}
}