We present ActioNet, an interactive end-to-end platform for data collection and augmentation of task-based dataset in a 3D environment. Using ActioNet, we collected a large-scale comprehensive task-based dataset, comprising over 3000 hierarchical task structures and videos. Using the hierarchical task structures, the videos are further augmented across 50 different scenes to give over 150,000 videos.
We created the ActioNet platform. Given an input comprising (a) task descriptions generated by annotators, using the (b) ActioNet’s GUI, the task-based dataset comprising (c) annotated videos and (d) hierarchical task structures can be collected. The (e) task-based dataset can be expanded using the (f) ActioNet’s data augmentation unit to generate the (g) augmented dataset.
Demonstration of ActioNet task-based dataset collection GU
Results
Example of video dataset of the annotated task(Left) Generated tasks. (Right) Original Tasks
BibTeX
@inproceedings{duan2020actionet,
title={Actionet: An Interactive End-To-End Platform For Task-Based Data Collection And
Augmentation In 3D Environment},
author={Duan, Jiafei and Yu, Samson and Tan, Hui Li and Tan, Cheston},
booktitle={2020 IEEE International Conference on Image Processing (ICIP)},
pages={1566--1570},
year={2020},
organization={IEEE}
}
Acknowledgement
This research was supported by NRF grant NRF2015-NRFISF001-2541 and by the Agency for Science, Technology and Research (A*STAR) under its AME Programmatic Funding Scheme (Project A18A2b0046)
Jiafei Duan 