Predictive Intention Recognition Dataset for Collaborative Assembly

Description

This dataset is part of the research work "Predictive intention recognition using deep learning for collaborative assembly" presented at CoDIT 2024 as part of the RICAIP project, funded by the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 857306. The dataset includes about a thousand images of workers engaged in different hand motions during some unspecified task. Each image has been annotated with the human hands, serving as the primary input for YOLACT hand segmentation. The folder at the bottom has the following format:

• Hand Segmentation Dataset: contains the entire set of images used to train the YOLACT model as well as their annotations.

• LSTM Dataset Sample: A small portion of the data used to train the LSTM model which we are not at liberty to publish due to project constraints. Contains short video samples and extracted frames with annotations.

Usage

This dataset is intended for research purposes, specifically for the development and testing of models aimed at predicting human intentions in collaborative environments. Researchers and developers can use this data to train and validate their models. Unblurred images of individuals appearing in this dataset should not be used in any publications or papers without permission.

If you use this dataset in your research, please cite our work as follows:

@inproceedings{rekik2024predictive,
  title={Predictive intention recognition using deep learning for collaborative assembly},
  author={Rekik, Khansa and Gajjar, Nishant and Silva, Grimaldo and M{\"u}ller, Rainer},
  booktitle={Int. Conf. on Control, Decision and Information Technologies (CoDIT)},
  year={2024},
  organization={IEEE}
}

Contact

For any questions or additional information, please contact:

• Khansa Rekik: k.rekik@zema.de
• Nishant Gajjar: nishantkgajjar@outlook.com
• Rainer Müller: rainer.mueller@zema.de
• Grimaldo Silva: jose.jgrimaldo@gmail.com

Download Link

• URL: https://dateiaustausch.zema.de/s/TioM9wJMZRsoM88
• Password: CoDit_2024

In case the link is temporarily offline, please, contact j.osejgrimaldo+dataset@gmail.com for a mirror link.