ZoomCap: Evidence-Routed Visual Zooming for Recoverable Visual Information Transfer

Key points

• Detailed captions transfer visual information into vision-language model supervision, but quality is easily conflated with length: long single-pass outputs still miss small objects, text, and spatial relations or promote uncertain observations into confident claims.1
• ZoomCap uses a draft caption to route conditional knowledge retrieval, OCR-sensitive reading, coverage-gap recovery, and adaptive visual zooming, allocating deeper crops only to the most complex eligible region, deriving spatial relations from grounded geometry, and fusing non-duplicate evidence.1
• We also introduce FactCycle, which scores caption-derived atomic claims on two blind anchors--source image (truthfulness) and reconstruction (realizability)--separating fabrication from generator loss claim by claim; human and multi-judge comparisons complement this fact-level view.1
• The evidence also transfers: in continual pre-training, ZoomCap-450K improves the 11-benchmark average by 1.31-2.38 points over the strongest competing corpus and Ling-Tiny-VL-v3 gains 1.4 points on a production document/OCR monitor suite (1.5 on perception) at 60B tokens; in medical pre-training, 2.64M ZoomCap-captioned images improve five VQA benchmarks by 3.04 points (vs. 1.76 for same-image Gemini-3.1-Pro captions).1

Sources (1)

• [1]ZoomCap: Evidence-Routed Visual Zooming for Recoverable Visual Information Transfer
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 06:10 AM

  We formulate fine-grained captioning as recoverable visual information transfer, in which a teacher acquires missing evidence under a finite observation budget and converts it into reusable supervision.

  Detailed captions transfer visual information into vision-language model supervision, but quality is easily conflated with length: long single-pass outputs still miss small objects, text, and spatial relations or promote uncertain observations into confident claims.

Extractive summary: sentences quoted from the sources.