Paper Overview
Field: Computer Vision Authors: Xiaoqi Zhao, Youwei Pang, Kailai Zhou arXiv: 2611.08486
Abstract (translated from Chinese summary)
Scientific figures give research ideas a visual form through choices of abstraction, composition, and emphasis. Generative AI offers new ways to realize these choices, raising questions about explanatory clarity and purposeful visual expression. We examine these questions through a visual retrospective of 371 motivation and framework figures from 200 papers. The corpus comprises a quota-based main sample of 160 papers spanning four temporal periods and a separate purposive style-contrast sample of 40 papers. A single multimodal judge evaluates six dimensions of communication and design, with three repeated judgments per figure. We attempt source-constrained revisions of all 371 figures, of which 369 pass the recorded fidelity checks. Original-figure ratings show no uniform decline across periods. Revisions improved the average clarity score by 0.428 points and reduced aesthetic consistency by 0.111 points (on a five-point scale). In the main sample, motivation figures gained more than framework figures in average scores across the six dimensions. Case analyses show that adding explanation can clarify contrasts, but issues of typographic hierarchy and visual economy remain unresolved. These model-based findings support researcher-led design, with AI assisting exploration, and authors retaining responsibility for explanatory purpose, composition, and scientific relationships.
Key Findings
- Corpus: 371 figures (motivation + framework) from 200 papers; 160-paper quota-based main sample across four periods plus 40-paper purposive style-contrast sample.
- Method: One multimodal judge rating six communication/design dimensions, three repeated judgments per figure; source-constrained AI revisions of all figures (369/371 passed fidelity checks).
- Results: No uniform decline in original-figure quality across time; AI revisions improved clarity (+0.428) but slightly reduced aesthetic consistency (-0.111) on a five-point scale; motivation figures benefited more than framework figures.
- Implication: AI can aid figure exploration, but authors should remain responsible for explanatory purpose, composition, and scientific accuracy.
*Auto-collected on 2026-11-07*