JamesWang007 / iros20-6d-pose-tracking

[IROS 2020] se(3)-TrackNet: Data-driven 6D Pose Tracking by Calibrating Image Residuals in Synthetic Domains

iros20-6d-pose-tracking

This is the official implementation of our paper "se(3)-TrackNet: Data-driven 6D Pose Tracking by Calibrating Image Residuals in Synthetic Domains" accepted in International Conference on Intelligent Robots and Systems (IROS) 2020. [PDF]

Abstract: Tracking the 6D pose of objects in video sequences is important for robot manipulation. This task, however, introduces multiple challenges: (i) robot manipulation involves significant occlusions; (ii) data and annotations are troublesome and difficult to collect for 6D poses, which complicates machine learning solutions, and (iii) incremental error drift often accumulates in long term tracking to necessitate re-initialization of the object's pose. This work proposes a data-driven optimization approach for long-term, 6D pose tracking. It aims to identify the optimal relative pose given the current RGB-D observation and a synthetic image conditioned on the previous best estimate and the object's model. The key contribution in this context is a novel neural network architecture, which appropriately disentangles the feature encoding to help reduce domain shift, and an effective 3D orientation representation via Lie Algebra. Consequently, even when the network is trained only with synthetic data can work effectively over real images. Comprehensive experiments over benchmarks - existing ones as well as a new dataset with significant occlusions related to object manipulation - show that the proposed approach achieves consistently robust estimates and outperforms alternatives, even though they have been trained with real images. The approach is also the most computationally efficient among the alternatives and achieves a tracking frequency of 90.9Hz.

Applications: model-based RL, manipulation, AR/VR, human-robot-interaction, automatic 6D pose labeling.

This repo can be used when you have the CAD model of the target object. When such model is not available, checkout our another repo BundleTrack, which can be instantly used for 6D pose tracking of novel unknown objects without needing CAD models

Bibtex

@conference {wense3tracknet,
	title = {se(3)-TrackNet: Data-driven 6D Pose Tracking by Calibrating Image Residuals in Synthetic Domains},
	booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
	year = {2020},
	month = {10/2020},
	url = {http://arxiv.org/abs/2007.13866},
	author = {Wen, Bowen and Mitash, Chaitanya and Ren, Baozhang and Bekris, Kostas}
}

(New) Application to visual feedback control

Some example experiments using se(3)-TrackNet in our recent work "Vision-driven Compliant Manipulation for Reliable, High-Precision Assembly Tasks", RSS 2021.

Supplementary Video:

Click to watch

Results on YCB

About YCBInEOAT Dataset

Due to the lack of suitable dataset about RGBD-based 6D pose tracking in robotic manipulation, a novel dataset is developed in this work. It has these key attributes:

• Real manipulation tasks

• 3 kinds of end-effectors

• 5 YCB objects

• 9 videos for evaluation, 7449 RGBD in total

• Ground-truth poses annotated for each frame

• Forward-kinematics recorded

• Camera extrinsic parameters calibrated

Link to download this dataset is provided below under 'Data Preparation'. Example manipulation sequence:

Current benchmark:

More details are in the paper and supplementary video.

Dependency

There are 2 options:

• Option 1: Create your anaconda environment by conda env create -f environment.yml. Then source activate bowen

• Option 2 (Recommended): Use docker and pull the pre-built image

   docker pull wenbowen123/se3_tracknet:latest
  

Data Download

1. YCB_Video dataset
2. data_organized (15G). It is the reorganized YCB_Video data for convenience. Then extract it under your YCB_Video dataset directory, e.g. YCB_Video_Dataset/data_organized/0048/
3. YCBInEOAT dataset (22G)
4. Our pretrained weights on YCB_Video and pretrained weights on YCBInEOAT
5. Our generated synthetic YCB_Video training data (~15G for each object) and synthetic YCBInEOAT trainnig data (~15G for each object)

6. se(3)-TrackNet's output pose estimations of YCB_Video and se(3)-TrackNet's output pose estimations of YCBInEOAT

Prediction on YCB_Video and YCBInEOAT

Please refer to predict.py and predict.sh

Benchmarking

Please refer to eval_ycb.py and eval_ycbineoat.py

Training

1. Edit the config.yml. Make sure the paths are correct. Other settings need not be changed in most cases.
2. Then python train.py