Realsee3D is a large-scale, multi-view RGB-D dataset containing 10,000 indoor scenes, comprising real-world residential scenes captured by 3D LiDAR Camera and procedurally generated scenes.
- Large Scale: 10,000 unique indoor scenes, comprising 95,962 rooms and 299,073 viewpoints/RGB-D pairs.
- Rich Data: Panoramic RGB-D captures with complete room-level coverage.
- Comprehensive Annotations: Includes pixel-wise semantic segmentation (available for both real-world and synthetic scenes), with CAD drawings, floor plans, 3D detection labels and more (forthcoming).
- Diverse Scenes: Comprising 1,000 real-world scenes with varied layouts and decoration styles, and 9,000 procedurally generated scenes utilizing over 100 designer-curated style templates, ensuring diverse furniture models and styles for robust training and testing.
The Realsee3D dataset is organized into individual scenes, each containing detailed multi-view RGB-D data.
- Data Structure & Usage: DATASET_STRUCTURE.md provides a comprehensive explanation of the file organization, data formats, and detailed usage instructions.
- Methodology: Visit our official website for more detailed information on how real-world data is collected and how synthetic data is generated.
To access the dataset, you must sign a Data Usage Agreement (PDF format). You can download the agreement directly here. Please send your request, specifying your intended use, to developer@realsee.com with the subject [Realsee3D Dataset Application]. Once your application is approved and you have signed the agreement, we will reply with download instructions and links.
Already have Phase I access? You do not need to submit a new agreement. Simply forward your previous approval email to developer@realsee.com, including a Google-Drive-accessible email address, and we will grant you download access to the Phase II data directly.
For a detailed breakdown of dataset statistics, please refer to metadata/README.md.
- 2026-06-25: Phase II data released — per-viewpoint pixel-wise semantic segmentation (
segment.png) for both real-world and synthetic scenes. Real-world maps are current-model predictions (not manual annotation); ground-truth annotations are planned for a future release. See DATASET_STRUCTURE.md for details. - 2025-12-05: Phase I data(RGB-D pano and extrinsics) released.
- 2025-11-28: Dataset introduction and official website release.
If you use the Realsee3D dataset in your research, please cite our paper:
@misc{Li2025realsee3d_data,
doi = {10.5281/zenodo.17826242},
url = {https://doi.org/10.5281/zenodo.17826242},
author = {Li, Linyuan and Wu, Yan and Li, Xi and Wang, Lingli and Rao, Tong and Zhou, Jie and Pan, Cihui and Hui, Xinchen},
title = {Realsee3D: A Large-Scale Multi-View RGB-D Dataset of Indoor Scenes (Version 1.1)},
publisher = {Zenodo},
year = {2026}
}
The accompanying code for data parsing, visualization, and evaluation is released under the MIT License, allowing for free use, modification, and distribution.