SynpickVP Dataset

This repository contains modules to load and preprocess the SynpickVP dataset

SynpickVP is a synthetic dataset for video-based prediction tasks in bin-picking scenarios. This dataset consisting of videos of various bin-picking scenarios in which a suction-cap gripper robot moves in arbitrary directions in a cluttered box containing different objects. SynpickVP is formed by processing a selection of from the Synpick dataset. Namely, all sequences in SynpickVP fulfill the following conditions:

• All objects are already placed in the box

• The robotic gripper is in the box

• The robotic gripper is moving in the sequence

                         Video                                    Semantic Segmentation                Instance Segmentation
  

Download

To download the data, please contact Angel Villar-Corrales.

Contact and Citation

This repository is maintained by Angel Villar-Corrales,

Please consider citing the following works if you find our work or our repository helpful.

@inproceedings{villar2022MSPred,
  title={MSPred: Video Prediction at Multiple Spatio-Temporal Scales with Hierarchical Recurrent Networks},
  author={Villar-Corrales, Angel and Karapetyan, Ani and Boltres, Andreas and Behnke, Sven},
  booktitle={British Machine Vision Conference (BMVC)},
  year={2022}
}

@inproceedings{periyasamy2021synpick,
  title={Synpick: A dataset for dynamic bin picking scene understanding},
  author={Periyasamy, Arul Selvam and Schwarz, Max and Behnke, Sven},
  booktitle={International Conference on Automation Science and Engineering (CASE)},
  year={2021},
}

In case of any questions or problems regarding the project or repository, do not hesitate to contact the authors at villar@ais.uni-bonn.de.