Deep Recognition of Vanishing-Point-Constrained Building Planes in Urban Street Views

By Zhiliang ZENG, Mengyang WU, Wei ZENG, and Chi-Wing Fu

[2021/03/06: updated demo code & pretrained model]

Introduction

This repository contains the inference code for our TIP paper: 'Deep Floor Plan Recognition Using a Multi-Task Network with Room-Boundary-Guided Attention'.

Requirements

Please install OpenCV
Please install Python
Please install tensorflow-gpu

Our origin code was written by using tensorflow-gpu==1.10.1 & Python2.7, now we tried to upgrade to tensorflow-gpu==1.14.0 & Python3.6. The code is not well tested, it may be a little different from the results shown in our paper.

Python packages

[numpy]
[scipy]
[Pillow]
[pylsd]
[matplotlib]

Vanishing point tool

We are using the method proposed by 'A-Contrario Horizon-First Vanishing Point Detection Using Second-Order Grouping Laws', you can find & download the Matlab tool here.

Data

Please send request email to: (zlzeng6@gmail.com)

Usage

To use our demo code, please first download the our pretrained model here and GeoNet here, unzip and put it into "pretrained" folder, then run

cd segmentation/
python infer.py

Citation

If you find our work useful in your research, please consider citing:

@ARTICLE{zlzeng2020deepbuildingplane,
  author={Z. {Zeng} and M. {Wu} and W. {Zeng} and C. -W. {Fu}},
  journal={IEEE Transactions on Image Processing}, 
  title={Deep Recognition of Vanishing-Point-Constrained Building Planes in Urban Street Views}, 
  year={2020},
  volume={29},
  number={},
  pages={5912-5923},
  doi={10.1109/TIP.2020.2986894}}

zlzeng / DeepBulidingPlanes