SeVeR: Selective Visual Exposure and Retrieval for 3D Medical Image Question Answering

arXiv:2608.25630v1 Announce Type: new Abstract: Volumetric medical VQA requires reasoning over long and redundant 3D visual token sequences, especially in multi-sequence MRI where complementary modalities provide diverse diagnostic cues but expose the decoder to many repeated anatomical regions. To investigate reasoning under multi-sequence visual redundancy, we first introduce BreMRIs-VQA, a clinically curated breast MRI benchmark with 1.19M QA pairs from 71.0K sequences and 12.9K patients, covering both free-text and multiple-choice questions. We further propose SeVeR, a selective visual exposure framework that compresses dense volumes into modality-wise prototypes and retrieves complementary multi-level evidence with change-aware gated attention during decoding, trained with a marginal-utility self-consistency objective that suppresses unhelpful retrieval. Experiments on BreMRIs-VQA and public benchmarks show that SeVeR improves both discriminative and generative performance while exposing substantially fewer visual tokens.

Related

• PathoArgus: Advancing Evidence-Grounded Long-Context Visual Reasoning across Gigapixel Whole-Slide and Multi-Slide Case Contexts
• UltraVR: A Diagnostic Ultra-Resolution Image-VQA Benchmark for Evidence-Grounded Reasoning
• MI-CXR: A Benchmark for Longitudinal Reasoning over Multi-Interval Chest X-rays

Source: arXiv cs.CV | 2026-08-27