MolGrapher: Graph-based Chemical Recognition
A graph-based deep learning approach for optical chemical structure recognition that outperforms image captioning …
Benchmark - Page 2
MolMole: Unified Vision Pipeline for Molecule Mining
A vision-based deep learning framework that unifies molecule detection, reaction parsing, and OCSR for page-level …
MolScribe: Image-to-Graph Molecular Recognition
Image-to-graph generation model for OCSR that predicts atoms, bonds, and coordinates jointly to better handle …
OCSU: Optical Chemical Structure Understanding
OCSU task for translating molecular images into multi-level descriptions. Introduces Vis-CheBI20 dataset and …
ABC-Net: Divide-and-Conquer SMILES Recognition
Deep learning OCSR model using keypoint estimation to detect atom and bond centers for graph-based molecular structure …
Handwritten Chemical Structure Recognition with RCGD
An end-to-end framework (RCGD) and unambiguous markup language (SSML) for recognizing complex handwritten chemical …
MolMiner: Deep Learning OCSR with YOLOv5 Detection
Deep learning OCSR tool using YOLOv5 and MobileNetV2 to extract machine-readable molecular structures from scientific …
SwinOCSR: Vision Transformers for Chemical OCR
Deep learning model using Swin Transformer and Focal Loss for OCSR, achieving 98.58% accuracy on synthetic benchmarks.
Review of OCSR Tools (2020)
Comprehensive review and benchmarking of 30 years of Optical Chemical Structure Recognition (OCSR) methods and tools.
Chemical Structure Reconstruction with chemoCR
A hybrid system combining pattern recognition and rule-based expert systems to reconstruct chemical structures from …
ChemReader at TREC 2011 Chemical IR Track
ChemReader OCR software evaluation on TREC 2011 Chemical IR campaign achieving 93% accuracy on image-to-structure task.
CLEF-IP 2012 Benchmark Overview
Overview of the CLEF-IP 2012 benchmarking lab focusing on patent passage retrieval, flowchart recognition, and chemical …