Two-Parameter Persistence for Images via Distance Transform

Abstract

The distance transform of a binary image is a classic tool in computer vision and it has been widely used in the field of Topological Data Analysis (TDA) to study porous media. A common practice is to convert grayscale images to binary ones to apply the distance transform. In this work, by considering the threshold decomposition of a grayscale image, we prove that threshold decomposition and distance transform together to formulate a two-parameter filtration. This would offer the TDA community a concrete example to apply multi-parameter persistence on digital image analysis. We demonstrate our method on the firn dataset.

Cite

Text

Hu et al. "Two-Parameter Persistence for Images via Distance Transform." IEEE/CVF International Conference on Computer Vision Workshops, 2021. doi:10.1109/ICCVW54120.2021.00464

Markdown

[Hu et al. "Two-Parameter Persistence for Images via Distance Transform." IEEE/CVF International Conference on Computer Vision Workshops, 2021.](https://mlanthology.org/iccvw/2021/hu2021iccvw-twoparameter/) doi:10.1109/ICCVW54120.2021.00464

BibTeX

@inproceedings{hu2021iccvw-twoparameter,
  title     = {{Two-Parameter Persistence for Images via Distance Transform}},
  author    = {Hu, Chuan-Shen and Lawson, Austin and Chung, Yu-Min and Keegan, Kaitlin},
  booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
  year      = {2021},
  pages     = {4159-4167},
  doi       = {10.1109/ICCVW54120.2021.00464},
  url       = {https://mlanthology.org/iccvw/2021/hu2021iccvw-twoparameter/}
}