This project implements a Parts Based Detector in C++, described in the following paper:
Yi Yang, Deva Ramanan, "Articulated Pose Estimation with Flexible Mixtures-of-Parts," CVPR 2011
This is an open source project, developed on open platforms. At this point I do not have the desire or resources to support non-compliant compilers such as MSVC. If you are using GCC or Clang and run into issues, be sure to let me know. If you are running MSVC, I can only offer you limited assistance (mostly due to lack of knowledge about MSVC idiosyncracies).
The project has the following dependencies:
- OpenCV REQUIRED (for image processing)
- CMake REQUIRED (for building)
- Doxygen OPTIONAL (for documentation)
- OpenMP OPTIONAL (for multithreading)
- ROS OPTIONAL (for publishing detections on a ROS topic)
- ECTO OPTIONAL (for building and ECTO cell)
The project can be built in one of two modes:
- A standalone binary (for testing functionality)
- A shared library (for use in existing applications)
To configure the project, set the options at the top of CMakeLists.txt To build the project, follow the normal cmake routine from the root folder:
- mkdir build
- cd build
- cmake ..
To run the detector, please consult the Mainpage of the docs, or src/demo.cpp. Both contain examples of how the detector can be initialised and run.
The learning code is currently only in Octave/Matlab. This is because the detector supports a number of learning schema, and porting all of these to C++ is not practical at this time. Please consult the README within the matlab/ directory for instructions on training a model
This package is developed and maintained by Hilton Bristow, Willow Garage