Home

Awesome

Code of paper "Learning to Parse Wireframes in Images of Man-Made Environments", CVPR 2018

Folder/fileDescription
juncFor training junction detector.
linepxFor training straight line pixel detector.
wireframe.pyGenerate line segments/wireframe from predicted junctions and line pixels.
evaluationEvaluation of junctions and wireframes.

Requirements

The code is written and tested in python3, please install all requirements in python3.

Prepare data

    unzip v1.1.zip
    unzip pointlines.zip
    unzip linemat.zip

Note: --json means you put the hype-parameters in junc/hypes/1.json.

Training

Testing

Evaluation

The code for evaluation is put in evaluation/junc and evaluation/wireframe. Expected junction and wireframe precision/recall curve is like

<figure class="half"> <img src="evaluation/junc/junc_1_16.png", width=400/> </figure> <figure class="half"> <img src="evaluation/wireframe/1_0.5_0.5.png", width=400/> </figure>

Visualize the result

For visualizing the result, we recommend generating an html file using dominate to visualize the result of different methods in columns.

Citation

@InProceedings{wireframe_cvpr18,
author = {Kun Huang and Yifan Wang and Zihan Zhou and Tianjiao Ding and Shenghua Gao and Yi Ma},
title = {Learning to Parse Wireframes in Images of Man-Made Environments},
booktitle = {CVPR},
month = {June},
year = {2018}
}

License

You can use this code/dataset for your research and other usages, following MIT License.