CausRem (Ablation Dataset): Causal Visual Artifacts (Shadows & Reflections)

Project Page · Paper (arXiv:2509.18538)

What it is.
This dataset is an ablation subset focused on causal visual artifacts—namely shadows and reflections associated with removable objects. It is intended only for evaluation/analysis of artifact handling and does not include any post-removal ground truth images.

No post-removal GT. This subset is designed to estimate and analyze causal artifacts; there are no “clean” (object-free) ground-truth images included.

---

Dataset Summary

• Content: Real-world images with human-annotated object masks and artifact masks (shadow or reflection).
• Counts: The exact number of images in this subset may vary. Please inspect the repository for the current totals:
  • Shadow cases: 213
  • Reflection cases: 255
• Use case: Ablation and evaluation of object-removal pipelines with respect to removing causal visual artifacts.

---

Disclaimer

The dataset consists of images collected from publicly available online sources and is provided for research and educational purposes only. We do not claim ownership of the images. If you are a copyright holder and wish your content to be removed, please contact us.