Edeg-Nerf

The implementation for the AUTOCON paper 3D wireframe model reconstruction of buildings from multi-view images using neural implicit fields

Abstract

The 3D wireframe model provides concise structural information for building reconstruction. Traditional geometry-based methods are prone to noise or missing data in 3D data. To address these issues, this paper introduces Edge-NeRF, a 3D wireframe reconstruction pipeline using neural implicit fields. By leveraging 2D multi-view images and their edge maps as supervision, it enables self-supervised extraction of 3D wireframes, thus eliminating the need for extensive training on large-scale ground-truth 3D wireframes. Edge-NeRF constructs neural radiance fields and neural edge fields to optimize scene appearance and edge structure simultaneously, and then the wireframe model is fitted from coarse to fine based on the extracted 3D edge points. Furthermore, a synthetic multi-view image dataset of buildings with 3D wireframe ground truth annotations is introduced. Experimental results demonstrate that Edge-NeRF outperforms other geometry-based methods in all evaluation metrics

Enviroments

This code has been tested with Ubuntu 22.04, one 3090 GPU with CUDA 11.8, Python 3.8, Pytorch 2.4.0, and Pytorch-lightning 1.9.5.

Dataset

We introduce BEMI dataset, which contains 30 different CAD models of buildings with wireframes selected from the RoofGraphDataset.

We render the multi-view image using Blender and detect 2D edges using a pre-trained PiDiNet.

Multi-view images and corresponding GT wireframes are available at Google Drive.

Usage

Extraction 3D Edge Points

Edeg-Nerf training process command:

python train.py \
--dataset_name blender --N_importance 64 --img_wh 400 400 
--noise_std 0 --num_epochs 6 --batch_size 1024 \
--optimizer adam --lr 5e-4 --lr_scheduler steplr \
--decay_step 2 4 8 --decay_gamma 0.5 \
--root_dir abstrction/wf001 --exp_name wf001

Extract Edges from occupied field command:

wireframe_fitting/extract_point_cloud.py.

Reconstruction of Wireframes

wireframe_fitting/wireframe_fitting.py.

Visualization and Evaluation

To visualize and evaluate extracted wireframes, run wireframe_fitting/visualize_and_eval.py

Install point_cloud_utils to perform evaluations on chamfer distance, precision, recall, F-score and IOU.

Citation

If our work is useful for your research, please consider citing:

@article{doi:10.1016/j.autcon.2025.106145,
author = {Weiwei Fan and Xinyi Liu and Yongjun Zhang and Dong Wei and Haoyu Guo and Dongdong Yue},
title = {3D wireframe model reconstruction of buildings from multi-view images using neural implicit fields},
journal = {Automation in Construction},
volume = {174},
pages = {106145},
year = {2025},
publisher = {Elsevier},
issn = {0926-5805}
}

Acknowledgements

• Thanks for the NEF, which helps us to quickly implement our ideas.
• PC2WF
• LC2WF
• Pytorch Chamfer Distance
• point_cloud_utils