CitySurfaces

City-Scale Semantic Segmentation of Sidewalk Materials

Segmentation of sidewalk surfaces from street-level images

Sustainable Cities and Society 2022

Sidewalk paving materials mapped in Chicago, Washington DC and Brooklyn — Paving materials classified from street-level images of Chicago, Washington DC and Brooklyn, none of them in the training data. Thicker lines mark segments whose dominant material is not concrete.

CitySurfaces is a framework that combines active learning and semantic segmentation to locate, delineate, and classify sidewalk paving materials from street-level images. Our framework adopts a recent high-performing semantic segmentation model (Tao et al., 2020), which uses hierarchical multi-scale attention combined with object-contextual representations.

For more information, see the GitHub project.

Publications

CitySurfaces: City-scale semantic segmentation of sidewalk materials

Maryam Hosseini, Fabio Miranda, Jianzhe Lin, Claudio T. Silva

Sustainable Cities and Society, 2022

Paper arXiv Code

How to cite

BibTeX

@article{hosseini2022citysurfaces,
  title = {CitySurfaces: City-scale semantic segmentation of sidewalk materials},
  volume = {79},
  doi = {10.1016/j.scs.2021.103630},
  journal = {Sustainable Cities and Society},
  publisher = {Elsevier BV},
  author = {Hosseini, Maryam and Miranda, Fabio and Lin, Jianzhe and Silva, Claudio T.},
  year = {2022},
  month = apr,
  pages = {103630},
  eprint = {2201.02260},
  archiveprefix = {arXiv}
}