Metraipolis › archive › dc_2c35f32d6a8a

Spot-the-shift: Evaluating Grounded Image Difference Captioning of Long-term Changes — arXiv:2609.10356

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.10356
License (as archived)
CC0-1.0
Archived
2026-10-16T11:00:03.054Z
SHA-256 of the archived text
ead97bb2c125f4696e3db885d09f6c6531eac3e0311ca2b1773ec68b83103dff
Stable URL of this page
https://metraipolis.com/documento/dc_2c35f32d6a8a

How to cite

Spot-the-shift: Evaluating Grounded Image Difference Captioning of Long-term Changes — arXiv:2609.10356 — arXiv (en), https://arxiv.org/abs/2609.10356 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2c35f32d6a8a (Metraipolis archive dc_2c35f32d6a8a, sha256 ead97bb2c125f4696e3db885d09f6c6531eac3e0311ca2b1773ec68b83103dff)
BibTeX
@misc{dc2c35f32d6a8a,
  title = {Spot-the-shift: Evaluating Grounded Image Difference Captioning of Long-term Changes — arXiv:2609.10356},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.10356},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2c35f32d6a8a; sha256 ead97bb2c125f4696e3db885d09f6c6531eac3e0311ca2b1773ec68b83103dff},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.10356. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-term change understanding from images of the same place revisited over time is a challenging task with applications in map maintenance and urban infrastructure monitoring. Prior work addresses it either through pixel-level prediction or difference captioning, neither of which is sufficient to reliably measure how well models detect and describe such changes. We introduce SPOT-THE-SHIFT, a human-verified benchmark for grounded image difference captioning of long-term changes in real-world driving scenes. Our benchmark provides natural language captions and spatial masks for structural changes across each image pair. We further propose an evaluation protocol that reliably assesses models' captioning ability, validated through human studies. Benchmarking state-of-the-art MLLMs, we find that models struggle with the fine-grained multi-image spatial capability required for this task. Finally, we develop a synthetic data generation pipeline that improves an off-the-shelf MLLM without sacrificing general capabilities.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-10 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.10356 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T10:02:34.384Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).