Publication Date

2025

Document Type

Dataset

Publisher

Edith Cowan University

School or Research Centre

School of Science

Description

This dataset was created for the evaluation of the EmbSCU method, suitable for solving the Scene Change Understanding (SCU) task. The SCU task involves predicting a changed location, describing a change, and generating language instructions for the robotic agent to revert a change. Current datasets, related to scene change understanding, can be divided into scene change detection (SCD) and image difference captioning (IDC) datasets. Unlike existing approaches, EmbSCU facilitates simultaneous change detection, description and language-based rearrangement instruction generation for the agent to revert changes. Although the EmbSCU dataset is simulated, it is highly complex, incorporating 104 unique indoor Ai2Thor rooms. EmbSCU encompasses 54 interactable indoor object categories.

DOI

10.25958/e7jp-p396

Methodology

Data was collected via Ai2Thor embodied Ai simulator.

Start of data collection time period

2022

End of data collection time period

2024

File Format(s)

txt, png, json

File Size

1.56 GB

Creative Commons License

Creative Commons Attribution-Noncommercial 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 4.0 License.

Contact

mariia.khan@ecu.edu.au

Download

Included in

Data Science Commons

Share

 
COinS