Metraipolis › archive › dc_bbb2c587786e

Think Before You Locate: Clinical Chain-of-Though Supervision for Sequential Medical Image Grounding — arXiv:2611.00041

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.00041
License (as archived): CC0-1.0
Archived: 2026-11-03T10:54:44.840Z
SHA-256 of the archived text: 14759246f5c8a9fb4d91c8d1b33238cfaa64b76fdf490f3cc66154b1bc6bb07e
Stable URL of this page: https://metraipolis.com/documento/dc_bbb2c587786e

How to cite

Think Before You Locate: Clinical Chain-of-Though Supervision for Sequential Medical Image Grounding — arXiv:2611.00041 — arXiv (en), https://arxiv.org/abs/2611.00041 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bbb2c587786e (Metraipolis archive dc_bbb2c587786e, sha256 14759246f5c8a9fb4d91c8d1b33238cfaa64b76fdf490f3cc66154b1bc6bb07e)

BibTeX

@misc{dcbbb2c587786e,
  title = {Think Before You Locate: Clinical Chain-of-Though Supervision for Sequential Medical Image Grounding — arXiv:2611.00041},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.00041},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bbb2c587786e; sha256 14759246f5c8a9fb4d91c8d1b33238cfaa64b76fdf490f3cc66154b1bc6bb07e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00041. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Clinicians rarely localize a finding from one image alone. They compare current and prior scans, follow structures across slices, and match regions across views or modalities. Yet vision-language models for sequential medical grounding usually learn only from bounding boxes. Boxes show where the target is, but they do not explain why that region in that image answers the question. We study whether explicit, image-grounded reasoning can improve localization, and which tasks benefit from it. We introduce method, whose core is MedChain, a corpus of 182,573 filtered rationales spanning eight sequential grounding tasks and ten imaging modalities. A medical VLM (FlemingVL-38B) first drafts a caption-like rationale, and a multimodal refiner (Qwen3.5-397B-A17B) then re-examines the images, question and annotated box to write a box-conditioned explanation with explicit cross-image reasoning. We fine-tune Qwen3.5 on these rationales with deliberative chain-of-thought supervision (dCoT-SFT). Across 4B, 9B, and 27B models, dCoT-SFT sets a new state of the art on MedSG-Bench. The 27B model reaches 77.24 mIoU and 85.20 [email protected], improving over the strongest prior model by 4.69 and 5.49 points. In a paired comparison using fixed 4B checkpoints, dCoT-SFT improves over answer-only SFT by 3.56 mIoU (95% CI 3.13--4.00). The gains mainly come from change detection and object tracking, while performance drops on three appearance-matching tasks. Replacing a model's own rationale with a rationale from another task reduces mIoU by 12.2 points, showing that localization depends strongly on the reasoning context. Finally, adding GRPO with an IoU reward after dCoT-SFT gives only small and task-dependent gains, which we analyze through controlled multi-seed development experiments.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.00041 — 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 2026-11-03T10:49:38.445Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).