Skip to content
Projects

jingwu2121/reflect3r has 121 GitHub stars and 1 forks. GitHot measured 0 stars in 24 hours, 0 in 7 days, and 0 in 15 days. Last checked 2026-11-15 15:47 UTC. Last push 2026-07-28.

How these counts are measured

Measured star growth

in 24h
0
in 7 days
0
in 15 days
0
in 1 month
Update delayed

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

About

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

AI

README.md

Original

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

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

Badge with Logo Project Page Hugging Face Dataset Badge with Logo

teaser

Abstract

Mirror reflections are common in everyday environments and can provide stereo information within a single capture, as the real and reflected virtual views are visible simultaneously. We exploit this property by treating the reflection as an auxiliary view and designing a transformation that constructs a physically valid virtual camera, allowing direct pixel-domain generation of the virtual view while adhering to the real-world imaging process. This enables a multi-view stereo setup from a single image, simplifying the imaging process, making it compatible with powerful feed-forward reconstruction models for generalizable and robust 3D reconstruction. To further exploit the geometric symmetry introduced by mirrors, we propose a symmetric-aware loss to refine pose estimation. Our framework also naturally extends to dynamic scenes, where each frame contains a mirror reflection, enabling efficient per-frame geometry recovery. For quantitative evaluation, we provide a fully customizable synthetic dataset of 16 Blender scenes, each with ground-truth point clouds and camera poses. Extensive experiments on real-world data and synthetic data are conducted to illustrate the effectiveness of our method.

✨ News

  • [28.07.2026] Fix the Croco model issues and added a docker env for the reproduction.
  • [18.03.2026] Our synthetic evaluation dataset can now be downloaded in Voxel51

citation Citation

If you find this code 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},
}

settings Installation

  1. Clone the repo
git clone --recursive https://github.com/jingwu2121/reflect3r
cd reflect3r
  1. Install the environment, we follow the DUSt3R environment, please also refer to DUSt3R's issue page if there is any environment issue.
conda create -n reflect3r python=3.11 
conda activate reflect3r 
conda install pytorch torchvision pytorch-cuda=12.1 -c pytorch -c nvidia  # use the correct version of cuda for your system
pip install -r requirements.txt
  1. Download the model weights for the mirror detection from here and put it under the weight folder. This is the guide of the mirror detection's original repo: here.

🐳 Using docker for easier reproduction

I also provide a docker image with all the dependencies installed for an easier reproduction.

docker pull jingwu2121/reflect3r:latest

docker run --gpus all -it --name reflect3r-cuda118 \ 
-w /workspace  \
  jingwu2121/reflect3r:latest /bin/bash

cd /workspace/reflect3r

# Run our example
python reflect3r_pipeline.py --input_image_path examples/example1.png

database Synthetic Evaluation Data

dataset previews

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

Download from Voxel51 Group Dataset

Please check the instruction here!

Installation

pip install -U fiftyone openexr

Usage


import fiftyone as fo
from huggingface_hub import snapshot_download


# Download the dataset snapshot to the current working directory

snapshot_download(
    repo_id="Voxel51/reflect3er", 
    local_dir=".", 
    repo_type="dataset"
    )



# Load dataset from current directory using FiftyOne's native format
dataset = fo.Dataset.from_dir(
    dataset_dir=".",  # Current directory contains the dataset files
    dataset_type=fo.types.FiftyOneDataset,  # Specify FiftyOne dataset format
    name="reflect3er"  # Assign a name to the dataset for identification
)

🤗 Download from HF

Download our synthetic evaluation data here.

Scripts for Modelling in Blender

  • Use data_toolkit/add_mirrored_cam.py to model a mirrored camera against the mirror plane in Blender's scripting tab.
  • Use data_toolkit/render_depth.py to render the depth, rgb images and camera information from Blender scripting.

Generate the ground-truth point cloud for evaluation:

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

Demo Example

We include 3 example images in the examples folder. Run this code for the reflect3r demo.

python reflect3r_pipeline.py --input_image_path examples/example1.png

The result will be saved to results/reconstructed_point_cloud.ply.

Acknowledgement

Zirui Wang is supported by an ARIA research gift grant from Meta Reality Lab.

Our Code is built upon:

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.