AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Axis-Aligned Document Dewarping

Chaoyun Wang, I-Chao Shen, Takeo Igarashi, Caigui Jiang

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add axis-aligned geometric constraint loss (AL) (ablation)UVDoc-based dewarping network · DocUNet benchmarkAAD: 0.124 -> 0.105Improved
add AL (ablation)UVDoc-based dewarping network · DIR300 benchmarkAAD: 0.100 -> 0.076Improved
add axis alignment preprocessing (AP) (ablation)UVDoc-based dewarping network · DocUNet benchmarkAAD: 0.124 -> 0.120No clear effect
add AP (ablation)UVDoc-based dewarping network · DIR300 benchmarkAAD: 0.100 -> 0.069Improved
add AL and AP together (ablation)UVDoc-based dewarping network · DocUNet benchmarkAAD: 0.124 -> 0.099Improved
add AL and AP together (ablation)UVDoc-based dewarping network · DIR300 benchmarkCER: 0.243 -> 0.132Improved
add AL (ablation)UVDoc-based dewarping network · DocUNet benchmarkCER: 0.167 -> 0.183Hurt
add AL (ablation)UVDoc-based dewarping network · DIR300 benchmarkCER: 0.243 -> 0.206Improved
add AP (ablation)UVDoc-based dewarping network · DIR300 benchmarkLD: 7.691 -> 4.303Improved
use full model instead of UVDocdocument dewarping · DocUNet benchmarkMS-SSIM: 0.545 vs 0.543No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy