doxa-friend/MolMole_Patent300
收藏Hugging Face2025-07-21 更新2025-08-09 收录
下载链接：
https://hf-mirror.com/datasets/doxa-friend/MolMole_Patent300
下载链接
链接失效反馈官方服务：
资源简介：
MolMole_Patent300是一个经过策划的基准数据集，旨在评估从整页专利图像中提取化学信息的模型。数据集包含300个完整的专利页面图像，以及对应的分子和文本边框、反应的注释信息。它支持端到端的化学信息提取评估，包括分子边框检测、反应解析和光学化学结构识别。
MolMole_Patent300 is a curated benchmark designed to evaluate models that extract chemical information from full-page patent images. The dataset consists of 300 full patent page images along with corresponding annotations for molecule and text bounding boxes, and reactions. It enables end-to-end evaluation of chemical information extraction, including molecule bounding box detection, reaction parsing, and optical chemical structure recognition (OCSR).
提供机构：
doxa-friend
