Neural-Networks

A deep learning method using EfficientNet and Transformer to convert hand-drawn chemical structures into SMILES codes, …

A 26B parameter multimodal LLM for chemistry, combining InternViT-6B and ChemLLM-20B for molecular structure …

Open-source OCSR platform combining Mask R-CNN segmentation and Transformer recognition, trained on 450M+ synthetic …

A Transformer-based OCSR model introducing dual-path modules (CGFE and SDGLA) to improve global context awareness and …

An improved encoder-decoder model (EfficientNetV2 + Transformer) for converting hand-drawn chemical structures into …

Deep learning model using improved SwinTransformer encoder and attention-based feature fusion to convert molecular …

Multi-modal transformer combining vision, text, and layout encoding to extract complex Markush structures from patent …

A deep learning model for Optical Chemical Structure Recognition (OCSR) using SwinV2 and GPT-2 to convert molecular …

A graph-based deep learning approach for optical chemical structure recognition that outperforms image captioning …

A vision-based deep learning framework that unifies molecule detection, reaction parsing, and OCSR for page-level …

Image-to-graph generation model for OCSR that predicts atoms, bonds, and coordinates jointly to better handle …

A three-stage OCSR framework using SMILES pretraining, auxiliary bond/coordinate tasks, and reinforcement learning to …