SeasonScapes: Learning Large-scale Re-lightable 3D Landscapes with Seasonal Variation from Sparse Webcams

Timo Kleger; Qi Ma; Deheng Zhang; Luc Van Gool; Danda Pani Paudel

SeasonScapes: Learning Large-scale Re-lightable 3D Landscapes with Seasonal Variation from Sparse Webcams

Computer Vision and Pattern Recognition 2026-05-12 v1

Authors: Timo Kleger , Qi Ma , Deheng Zhang , Luc Van Gool , Danda Pani Paudel

Abstract

We introduce SeasonScapes framework and a the SeasonScapes dataset: Swiss Sparse-view Mountain Scenes with Seasonal Changes that covers over 50 km x 60 km, composed of more than 85,000 webcam images captured from 32 different locations across 13 timestamps throughout a full year. By projecting these timestamp-specific images onto a 3D mesh, we construct seasonal 3D landscapes that reflect natural appearance changes over time. To address occlusions and missing data, we leverage conditional diffusion models for image-guided inpainting directly on the mesh. The resulting completed meshes can be further relighted using standard physically-based renderer.

Keywords

3D reconstruction seasonal variation sparse view webcam imagery conditional diffusion mesh inpainting relighting

Cite

@article{arxiv.2605.09039,
  title  = {SeasonScapes: Learning Large-scale Re-lightable 3D Landscapes with Seasonal Variation from Sparse Webcams},
  author = {Timo Kleger and Qi Ma and Deheng Zhang and Luc Van Gool and Danda Pani Paudel},
  journal= {arXiv preprint arXiv:2605.09039},
  year   = {2026}
}

SeasonScapes: Learning Large-scale Re-lightable 3D Landscapes with Seasonal Variation from Sparse Webcams

Abstract

Keywords

Cite

Related papers