AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

WordScape: a Pipeline to extract multilingual, visually rich Documents with Layout Annotations from Web Crawl Data

Maurice Weber, Carlo Siebenschuh, Rory Marshall Butler, Anton Alexandrov, Valdemar Ragnar Thanner, Georgios Tsolakis, Haris Jabbar, Ian Foster, Bo Li, Rick Stevens, Ce Zhang

NeurIPS 2023 · Vision-language and multimodal › Human and embodied modeling · approach: Layout-aware document understanding

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
Pretrain on 10k WordScape samples before finetuning with 25 FUNSD samplesFaster R-CNN · FUNSDF1 @IoU 0.5: 0.621 -> 0.840Improved
Pretrain on 50k WordScape samples before finetuning with 50 FUNSD samplesFaster R-CNN · FUNSDF1 @IoU 0.5: 0.690 -> 0.870Improved
Pretrain on 50k WordScape samples before finetuning with 100 FUNSD samplesFaster R-CNN · FUNSDF1 @IoU 0.5: 0.723 -> 0.857Improved
Pretrain on 100k WordScape samples before finetuning with 149 FUNSD samplesFaster R-CNN · FUNSDF1 @IoU 0.5: 0.772 -> 0.882Improved
Pretrain on 5k WordScape samples before finetuning with 75 cTDaR samplesYOLOv8m · ICDAR 2019 cTDaR modernmAP @ IoU [0.50:0.95]: 0.869 ± 0.008 -> 0.924 ± 0.007Improved
Pretrain on 10k WordScape samples before finetuning with 150 cTDaR samplesYOLOv8m · ICDAR 2019 cTDaR modernmAP @ IoU [0.50:0.95]: 0.888 ± 0.011 -> 0.931 ± 0.010Improved
Pretrain on 10k WordScape samples before finetuning with 300 cTDaR samplesYOLOv8m · ICDAR 2019 cTDaR modernmAP @ IoU [0.50:0.95]: 0.949 ± 0.006 -> 0.961 ± 0.005Improved
Pretrain on 10k WordScape samples before finetuning with 600 cTDaR samplesYOLOv8m · ICDAR 2019 cTDaR modernmAP @ IoU [0.50:0.95]: 0.974 ± 0.003 -> 0.975 ± 0.003No clear effect
Pretrain on 200k WordScape images rather than random initializationYOLOv5 · DocLayNet, N_f=1kmAP @ IoU [0.50:0.95]: 0.299 -> 0.508Improved
Pretrain on 200k WordScape images rather than PubLayNetYOLOv5 · DocLayNet, N_f=1kmAP @ IoU [0.50:0.95]: 0.467 -> 0.508Improved
Pretrain on 200k WordScape images rather than PubLayNetYOLOv5 · DocLayNet, N_f=5kmAP @ IoU [0.50:0.95]: 0.659 -> 0.679Improved
Pretrain on 200k WordScape images rather than PubLayNetYOLOv5 · DocLayNet, N_f=20kmAP @ IoU [0.50:0.95]: 0.720 -> 0.734Improved

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