CaptainCook4D: A Dataset for Understanding Errors in Procedural Activities
(This page is under continuous update and construction.)
DISCLAIMER: The proposed data splits showcase the ways in which our dataset can be used. We do not claim that these splits are the best or the only way to use the dataset. We encourage the community to explore other ways to use the dataset and share their findings with us.
Update
Dec 2023: Released dataset version 1.0
Overview
Annotations
Task Graphs
https://github.com/user-attachments/assets/f32e37c6-adb9-4bd8-952b-7effa3de4b01
Step/Coarse Grained Action Annotations Overview
Data Collection and Annotation Illustration
https://github.com/user-attachments/assets/87cae353-536b-48ad-942e-5eeb7c725089
Error Categories
Dataset Statistics
Normal/Error Recording Statistics
Error Recipe Count
License & Consent
Our dataset is licensed under the Apache license 2.0: License.
Our dataset is approved by the Institutional Review Board (IRB) at the University of Florida: IRB Approval
All participants provided written consent for the data collection: Consent
Citation
If you have found our work to be useful in your research, please leave us a star and consider citing our work:
@misc{peddi2023captaincook4d,
title={{CaptainCook4D: A dataset for understanding errors in procedural activities}},
author={Rohith Peddi and Shivvrat Arya and Bharath Challa and Likhitha Pallapothula and Akshay Vyas and Jikai Wang and Qifan Zhang and Vasundhara Komaragiri and Eric Ragan and Nicholas Ruozzi and Yu Xiang and Vibhav Gogate},
year={2023},
eprint={2312.14556},
archivePrefix={arXiv},
primaryClass={cs.CV}
}