pyni / ea_repose

EA_RePOSE

This is the code for our submitted paper to ICRA 2023:

EA-Repose: Efficient and Accurate Feature-metric-based 6D Object Pose Refinement

Prerequisites

• Python >= 3.6
• Pytorch == 1.9.0
• Torchvision == 0.10.0
• CUDA == 10.1

Downloads

The dataset parts are from the Downloads part in RePOSE(https://github.com/sh8/RePOSE.git)

Our training results w/o dqn can be downloaded from this part

$ ROOT=/path/to/EA_RePOSE
$ mkdir  $ROOT/bestresult

Then copy the training weights in it

Dqn results can be downloaded from this part

$ ROOT=/path/to/EA_RePOSE
$ mkdir  $ROOT/bestresult_dqn

Then copy the training weights in it

Installation

1. Set up the python environment:

   $ pip install torch==1.9.0 torchvision==0.10.0
   $ pip install Cython==0.29.17
   $ sudo apt-get install libglfw3-dev libglfw3
   $ pip install -r requirements.txt
   
   # Install Differentiable Renderer
   $ cd renderer
   $ python3 setup.py install
   

2. Compile cuda extensions under lib/csrc:

   ROOT=/path/to/EA_RePOSE
   cd $ROOT/lib/csrc
   export CUDA_HOME="/usr/local/cuda-10.1"
   cd ../camera_jacobian
   python setup.py build_ext --inplace
   cd ../nn
   python setup.py build_ext --inplace
   cd ../fps
   python setup.py
   

3. Set up datasets:

   $ ROOT=/path/to/EA_RePOSE
   $ cd $ROOT/data
   
   $ ln -s /path/to/linemod linemod
   $ ln -s /path/to/linemod_orig linemod_orig
   
   $ cd $ROOT/cache/LinemodTest
   $ unzip ape.zip benchvise.zip .... phone.zip
   
   

Testing w/o DQN

Evaluate the ADD(-S) score

1. Generate the annotation data:

   python run.py --type linemod cls_type ape model ape
   

2. Test (The method of initial poses can be modified in configs/linemod.yaml):

   # Test on the LineMOD dataset with results of PVNET
   $ python run.py --type evaluate --cfg_file configs/linemod.yaml cls_type ape model ape mode PVNET
   
   # Test on the LineMOD dataset with results of PoseCNN
   $ python run.py --type evaluate --cfg_file configs/linemod.yaml cls_type ape model ape mode PoseCNN
   
   

3. TensorRT version:

Please convert weight file('.pth') into tensorRT file according to TensorRT.

Then change branch into main_tensorrt:

git checkout main_tensorrt

Then run the above orders again.

Testing with DQN

Evaluate the ADD(-S) score

1. Change branch into main_dqn:

git checkout main_dqn

2. Test (The method of initial poses can be modified in configs/linemod.yaml):

   # Test on the LineMOD dataset with results of PVNET
   $ python run.py --type dqn --cfg_file configs/linemod.yaml cls_type ape model ape mode PVNET
   
   # Test on the LineMOD dataset with results of PoseCNN
   $ python run.py --type dqn --cfg_file configs/linemod.yaml cls_type ape model ape mode PoseCNN
   
   

3. TensorRT version:

Then change branch into main_dqn_tensorrt:

git checkout main_dqn_tensorrt

Then run the above orders again.

Acknowledgement

Our code is largely based on Repose and PVNET. Thanks for their sharing.