Learning Joint Reconstruction of Hands and Manipulated Objects - ManoGrasp

Porting the MANO hand model to GraspIt! simulator

Yana Hasson, Gül Varol, Dimitris Tzionas, Igor Kalevatykh, Michael J. Black, Ivan Laptev, Cordelia Schmid, CVPR 2019

• Project page

Install

Setup ROS interface

This package uses a ROS interface for the GraspIt! simulator.

To install and setup this interface follow the instructions at https://github.com/graspit-simulator/graspit_interface.

Install package

git clone https://github.com/lwohlhart/mano_grasp.git
cd mano_grasp
python setup.py install --user --graspit_dir=$GRASPIT

Model

Model ManoHand will be automatically copied to $GRASPIT directory during the installation.

To copy a model without the code installation use the command:

python setup.py --copy_model_only --graspit_dir=$GRASPIT

Prepare objects

Make sure you have meshlab installed:

sudo apt install meshlab

To prepare object files (.obj, .stl, .ply, .off) for graspit:

python -m mano_grasp.prepare_objects --models_folder /PATH/TO/YOURDATASET/ --file_out YOURDATASET_objects.txt

Usually you want to apply some scaling to the objects to fit the hand, therefore append scales options:

--scales 1000

Use

python -m mano_grasp.prepare_objects --help

to see all available options.

Generate grasps

Start ROS master in one terminal:

roscore

Then in a second terminal start generator:

python -m mano_grasp.generate_grasps --models_file YOURDATASET_objects.txt --path_out PATH_TO_DATASET

Use

python -m mano_grasp.generate_grasps --help

to see all available options.

Citations

If you find this code useful for your research, consider citing:

@INPROCEEDINGS{hasson19_obman,
  title     = {Learning joint reconstruction of hands and manipulated objects},
  author    = {Hasson, Yana and Varol, G{\"u}l and Tzionas, Dimitris and Kalevatykh, Igor and Black, Michael J. and Laptev, Ivan and Schmid, Cordelia},
  booktitle = {CVPR},
  year      = {2019}
}