CylinderDepth

CylinderDepth: Cylindrical Spatial Attention for Multi-View Consistent Self-Supervised Surround Depth Estimation

Project Page | Paper | Supplementary Material | Code

CylinderDepth is a geometry-guided method for calibrated, time-synchronized multi-camera rigs that predicts dense metric depth. It addresses depth inconsistency across views by extending the receptive field and applying a cylindrical spatial attention mechanism that aggregates features across images according to their distances on a shared cylinder.

Citation

If you find this work useful, please cite our paper:

@InProceedings{Abualhanud2026,
  author    = {Abualhanud, Samer and Grannemann, Christian and Mehltretter, Max},
  booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops},
  title     = {CylinderDepth: Cylindrical Spatial Attention for Multi-View Consistent Self-Supervised Surround Depth Estimation},
  year      = {2026},
  note     = {Accepted for publication.}
}