ML Anthology

A Phase Field Variational Model with Arctangent Regularization for Saliency Detection

Meng Li, Xing Liu, Liming Tang
WACVW 2017 pp. 29-35
doi:10.1109/WACVW.2017.12 /wacvw/2017/li2017wacvw-phase/

Abstract

In this paper, we propose a phase field variational model with arctangent regularization for saliency detection. The classical Cahn-Hilliard model is used to extract features in complex image domain with highly anisotropic interfacial energy. Various visual attention features can be detected by minimizing the energy functional. The processing of saliency detection can be seen as a dynamical competition between saliency and background. Our method is task-dependent and easy to implement. Experimental results on various images show that our model successfully suppresses the background, while preserves the fine feature, like point, silhouette and texture very well, which is important in terms of human visual perception.

WACVW Semantic Scholar

Cite

Text

Li et al. "A Phase Field Variational Model with Arctangent Regularization for Saliency Detection." IEEE/CVF Winter Conference on Applications of Computer Vision Workshops, 2017. doi:10.1109/WACVW.2017.12

Markdown

[Li et al. "A Phase Field Variational Model with Arctangent Regularization for Saliency Detection." IEEE/CVF Winter Conference on Applications of Computer Vision Workshops, 2017.](https://mlanthology.org/wacvw/2017/li2017wacvw-phase/) doi:10.1109/WACVW.2017.12

BibTeX

@inproceedings{li2017wacvw-phase,
  title     = {{A Phase Field Variational Model with Arctangent Regularization for Saliency Detection}},
  author    = {Li, Meng and Liu, Xing and Tang, Liming},
  booktitle = {IEEE/CVF Winter Conference on Applications of Computer Vision Workshops},
  year      = {2017},
  pages     = {29-35},
  doi       = {10.1109/WACVW.2017.12},
  url       = {https://mlanthology.org/wacvw/2017/li2017wacvw-phase/}
}