[3DV 2026] Reflect3r: Single-View 3D Stereo Reconstruction Aided by Mirror Reflections

Jing Wu, Zirui Wang, Iro Laina, Victor Adrian Prisacariu
University of Oxford

Citation

If you find this data or find the paper useful for your research, please consider citing:

@article{wu2026reflect3r,
author = {Wu, Jing and Wang, Zirui and Laina, Iro and Prisacariu, Victor},
title = {{Reflect3r: Single-View 3D Stereo Reconstruction Aided by Mirror Reflections}},
journal = {3DV},
year = {2026},
}

Synthetic Data for Evaluation

This repo includes the synthetic data used in our paper: Reflect3r.

We built a collection of synthetic Blender scenes containing mirror reflections to evaluate single-view stereo reconstruction in the context of mirror.

Download our synthetic evaluation data here.

Download the original blender scenes
Download the processed data

Scripts for Modelling in Blender

Use this python script to model a mirrored camera against the mirror plane in Blender's scripting tab.
Use this python script to render the depth, rgb images and camera information from Blender scripting.

Generate the ground-truth point cloud for evaluation using this script:

python syn_gt_point_cloud_gen.py --scene_name '<SCENE TO PROCESS>' --save_root '/path/to/save'

Acknowledgement

The 16 Blender scenes are collected from online websites, we rearranged and cleaned the scenes and modelled a mirror on top of them for research purpose. Here we listed all the original download links for these scenes, we thank these designers for their great work.

Datasets:

jinggogogo
/

reflect3r_synthetic_data

[3DV 2026] Reflect3r: Single-View 3D Stereo Reconstruction Aided by Mirror Reflections

Citation

Synthetic Data for Evaluation

Acknowledgement

Paper for jinggogogo/reflect3r_synthetic_data

Reflect3r: Single-View 3D Stereo Reconstruction Aided by Mirror Reflections

image imagewidth (px) 114 8.57k