Mobile Robot Oriented Large-Scale Indoor Dataset for Dynamic Scene Understanding

THUD Robotic Dataset

.

Yi-Fan Tang†, Cong Tai†, Fang-Xing Chen†, Wan-Ting Zhang, Tao Zhang, Xue-Ping Liu, Yong-Jin Liu, Long Zeng∗

Tsinghua university

Most existing robotic datasets capture static scene data and thus are limited in evaluating robots’ dynamic perfor- mance. To address this, we present a mobile robot oriented large- scale indoor dataset, denoted as THUD (Tsinghua University Dynamic) robotic dataset, for training and evaluating their dynamic scene understanding algorithms. Specifically, the THUD dataset construction is first detailed, including organization, acquisition, and annotation methods. It comprises both real- world and synthetic data, collected with a real robot platform and a physical simulation platform, respectively.

Dataset Application Form PAPER Github

large scale & dynamic

Our dataset provides dynamic annotated data for large scale indoor scene which contains amount of dynamic objects that present significant challenges for robot tasks.

scene understanding

Our dataset supports training and testing for various robotic scene understanding tasks (object detection, semantic segmentation, robot relocalization, scene reconstruction, etc.)

selective focus

Our dataset contains both real and synthetic annotated data, the expansion of its size and capabilities has great potential in the future.

Rich labels

Multiple labels such as instance segmentation, semantic segmentation, 3D/2D object detection, Depth, RGB, pose, etc., widely applicable in various fields.

Display Video:

THUD-Robot Dataset:

Image 1

Mobile robot synthetic data acquisition platform:

Image 1

Statistics of annotations in our dataset:

Image 1

Synthetic Scenes:

Image 1

Real Scenes:

Image 1

RGB-D Datasets Comparison:

Image 1

Cooperative unit

Thank you to the following units for their support and assistance.

.

If you use the ScanNet data or code please cite:

@inproceedings{2024ICRA,
title={Mobile Oriented Large-Scale Indoor Dataset for Dynamic Scene Understanding},
author={Yi-Fan Tang, Cong Tai, Fang-Xin Chen, Wan-Ting Zhang, Tao Zhang, Yong-Jin Liu, Long Zeng*},
booktitle = {Mobile Oriented Large-Scale Indoor Dataset for Dynamic Scene Understanding, submitted to IEEE International Conference Robotic and Automation, 2024.}},

full papers