AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

CEDe: A collection of expert-curated datasets with atom-level entity annotations for Optical Chemical Structure Recognition

Rodrigo Hormazabal, Changyoung Park, Soonyoung Lee, Sehui Han, Yeonsik Jo, Jaewan Lee, Ahra Jo, Seung Hwan Kim, Jaegul Choo, Moontae Lee, Honglak Lee

NeurIPS 2022 · Detection and segmentation › Open-set and few-shot detection · approach: Industrial anomaly detection · 0 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 synthetic pretraining (1M images) before real data (ablation)CNN+GRU image-to-SMILES · CEDe UOBgraph matching accuracy (%): 5.12 -> 32.01Improved
add synthetic pretraining (10K images) before real data (ablation)FasterRCNN+Instance-pair Transformer · CEDe UOBgraph matching accuracy (%): 49.77 -> 67.94Improved
fine-tune on 20% of real data after synthetic pretraining (ablation)FasterRCNN+Instance-pair Transformer · CEDe UOBgraph matching accuracy (%): 67.94 -> 89.26Improved
fine-tune on 20% of real data after synthetic pretraining (ablation)CNN+GRU · CEDe JPOgraph matching accuracy (%): 15.28 -> 33.61Improved
fine-tune on real data after synthetic pretraining (ablation)CNN+GRU · CEDe UOBgraph matching accuracy (%): 32.01 -> 62.57Improved
use chemical entity detection with 10K data instead of CNN+GRU with 1M dataFasterRCNN+bond bbox · CEDe synthetic testgraph matching accuracy (%): 61.2 -> 85.92Improved
use FasterRCNN+bond bbox detection instead of rule-based ImagoFasterRCNN+bond bbox · CEDe JPO fine-tunedgraph matching accuracy (%): 41.39 -> 74.12Improved
use FasterRCNN+IPT instead of OSRAFasterRCNN+Instance-pair Transformer · CEDe CLEF fine-tunedgraph matching accuracy (%): 92.84 -> 81.64Hurt
use image-to-SMILES CNN+Transformer instead of detection with synthetic pretrainingCNN+Transformer · CEDe USPTO synthetic onlygraph matching accuracy (%): 75.43 -> 42.60Hurt
fine-tune CNN+GRU+Attention on real data (ablation)CNN+GRU+Attention · CEDe CLEFgraph matching accuracy (%): 23.18 -> 39.19Improved
fine-tune CNN+GRU+Attention on real data (ablation)CNN+GRU+Attention · CEDe JPOgraph matching accuracy (%): 17.22 -> 30.28Improved
fine-tune FasterRCNN+bond bbox on real data (ablation)FasterRCNN+bond bbox · CEDe JPOgraph matching accuracy (%): 49.44 -> 74.12Improved

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