Timezone: »

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

Tue Nov 29 09:00 AM -- 11:00 AM (PST) @ Hall J #1017

Optical Chemical Structure Recognition (OCSR) deals with the translation from chemical images to molecular structures, this being the main way chemical compounds are depicted in scientific documents. Traditionally, rule-based methods have followed a framework based on the detection of chemical entities, such as atoms and bonds, followed by a compound structure reconstruction step. Recently, neural architectures analog to image captioning have been explored to solve this task, yet they still show to be data inefficient, using millions of examples just to show performances comparable with traditional methods. Looking to motivate and benchmark new approaches based on atomic-level entities detection and graph reconstruction, we present CEDe, a unique collection of chemical entity bounding boxes manually curated by experts for scientific literature datasets. These annotations combine to more than 700,000 chemical entity bounding boxes with the necessary information for structure reconstruction. Also, a large synthetic dataset containing one million molecular images and annotations is released in order to explore transfer-learning techniques that could help these architectures perform better under low-data regimes. Benchmarks show that detection-reconstruction based models can achieve performances on par with or better than image captioning-like models, even with 100x fewer training examples.

Author Information

Rodrigo Hormazabal (LG AI Research)
Changyoung Park (LG AI Research)
Soonyoung Lee (LG AI Research)

Soonyoung Lee is a member of Vision lab. in LG AI Research. He received BS, MS and Ph.D degree in Electrical Engineering at Seoul National Univ. His research focused on a correspondence matching between different visual modalities. Recently, he is working on object detection and graph neural network for document understanding.

Sehui Han (Seoul National University)
Yeonsik Jo (LG AI research)
Jaewan Lee (Korea Advanced Institute of Science & Technology)
Ahra Jo (Ulsan National Institute of Science and Technology)
Seung Hwan Kim (LG AI Research)
Jaegul Choo (Korea Advanced Institute of Science and Technology)
Moontae Lee (University of Illinois at Chicago)
Honglak Lee (LG AI Research / U. Michigan)

More from the Same Authors