Preprint
Aug 2026
VizAnchor: Decoding Manipulation Intent from Tampering Visualizations via Dual-Anchor Reasoning
VizAnchor is proposed, a framework for visualization manipulation understanding through dual-anchor evidence construction and VLM-based reasoning that accurately localizes manipulations and produces faithful explanations of their manipulation, misleaders, and misleading intents.
Xiaotian Zhang, Hua-Yuan Ye, Hai-Yang Zhang et al.
· 0 citations