Harmonious Semantic Line Detection via Maximal Weight Clique Selection

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

CVPR 2021 pp. 16737-16745

doi:10.1109/CVPR46437.2021.01646 /cvpr/2021/jin2021cvpr-harmonious/

Abstract

A novel algorithm to detect an optimal set of semantic lines is proposed in this work. We develop two networks: selection network (S-Net) and harmonization network (H-Net). First, S-Net computes the probabilities and offsets of line candidates. Second, we filter out irrelevant lines through a selection-and-removal process. Third, we construct a complete graph, whose edge weights are computed by H-Net. Finally, we determine a maximal weight clique representing an optimal set of semantic lines. Moreover, to assess the overall harmony of detected lines, we propose a novel metric, called HIoU. Experimental results demonstrate that the proposed algorithm can detect harmonious semantic lines effectively and efficiently. Our codes are available at https://github.com/dongkwonjin/Semantic-Line-MWCS.

PDF CVPR Semantic Scholar

Cite

Text

Jin et al. "Harmonious Semantic Line Detection via Maximal Weight Clique Selection." Conference on Computer Vision and Pattern Recognition, 2021. doi:10.1109/CVPR46437.2021.01646

Markdown

[Jin et al. "Harmonious Semantic Line Detection via Maximal Weight Clique Selection." Conference on Computer Vision and Pattern Recognition, 2021.](https://mlanthology.org/cvpr/2021/jin2021cvpr-harmonious/) doi:10.1109/CVPR46437.2021.01646

BibTeX

@inproceedings{jin2021cvpr-harmonious,
  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 = {Conference on Computer Vision and Pattern Recognition},
  year      = {2021},
  pages     = {16737-16745},
  doi       = {10.1109/CVPR46437.2021.01646},
  url       = {https://mlanthology.org/cvpr/2021/jin2021cvpr-harmonious/}
}