[CVPR 2021] Harmonious Semantic Line Detection via Maximal Weight Clique Selection

Dongkwon Jin, Wonhui Park, Seong-Gyun Jeong, and Chang-Su Kim

Official implementation for "Harmonious Semantic Line Detection via Maximal Weight Clique Selection" [paper] [supp] [video] [arxiv].

Video

Requirements

• PyTorch >= 1.3.1
• CUDA >= 10.0
• CuDNN >= 7.6.5
• python >= 3.6

Installation

Create conda environment:

    $ conda create -n MWCS python=3.6 anaconda
    $ conda activate MWCS
    $ pip install opencv-python==3.4.2.16
    $ conda install pytorch==1.6.0 torchvision==0.7.0 cudatoolkit=10.1 -c pytorch

Download repository:

    $ git clone https://github.com/dongkwonjin/Semantic-Line-MWCS.git

Instruction

1. Download preprocessed data for SEL, SEL_Hard, and NKL(SL5K) datasets to root/Preprocessing/. You can generate these data using the source codes in Preprocessing/. SEL and SEL_Hard datasets are provided in here. NKL dataset is provided in here.

2. Download our model parameters to root/Modeling/ if you want to get the performance of the paper.

3. Edit config.py. Please modify settings for path in the script file. Also, if you want to get the performance of the paper, please input run_mode to 'test_paper'.

4. Run with

cd Semantic-Line-MWCS-master/(Modeling or Prerpocessing)/(..)/code/
python main.py

Reference

@Inproceedings{
    Jin2021MWCS,
    title={Harmonious Semantic Line Detection via Maximal Weight Clique Selection},
    author={Jin, Dongkwon and Park, Wonhui and Jeong, Seong-Gyun and Kim, Chang-Su},
    booktitle={CVPR},
    year={2021}
}