VERDICT：在真实文档的光学化学结构识别（OCSR）中，一致性优于像素空间验证

中文摘要 AI 辅助

光学化学结构识别（Optical Chemical Structure Recognition, OCSR）将已发表文献中的二维分子图示转换为SMILES格式，对构建大规模化学训练数据集愈发重要。该规模的自动化需在无真实标注的情况下识别不可靠预测。研究在263幅带有验证真实标注的ACS期刊图示上，对比了三类无标注信号：模型置信度、重渲染相似度及识别器间的一致性。像素空间重渲染的表现仅略优于随机（AUROC为0.547，95%置信区间[0.465,0.629]），对其采用神调优阈值后，每幅图像的正确标注从0.745降至0.205；而四个架构各异的识别器间的一致性达到了0.916的AUROC（[0.880,0.952]）。四取二规则接受81.7%的图像，精度为88.8%；四取三规则接受52.1%的图像，精度为98.5%。在CLEF-IP、UOB和USPTO数据集上呈现相同规律，该差异在合成基准中被掩盖，因合成基准中重渲染预测自然与输入相似。物质过滤器移除了2193个通配符和R基团片段的虚假一致性，之后四取三规则拒绝了全部68个通用图示。VERDICT随后应用于PMC开放获取数据集，为4833个分子生成了6146个结构标注；化学家对两个独立样本中的400个发布标注进行裁定，得出四取三层的精度为0.995，四取二层的精度为0.958。因此VERDICT可生成经验证的标注，用于连接结构图像、机器可读表示与来源出版物信息的多模态分子数据库；在SES AI的Molecular Universe平台中，VERDICT还充当了搜索和检索分子记录的基于图像的接口。

英文摘要

Optical Chemical Structure Recognition (OCSR) converts 2D molecular depictions in the published literature into SMILES, and is increasingly important for constructing large-scale chemical training datasets. Automation at that scale requires identifying unreliable predictions in the absence of ground truth. Three families of label-free signals were compared on $263$ ACS journal depictions with verified ground truth: model confidence, re-rendering similarity, and agreement among recognizers. Pixel-space re-rendering performed little better than chance (AUROC $0.547$, $95\%$ CI $[0.465,0.629]$), and an oracle-tuned threshold on it reduced correct labels per image from $0.745$ to $0.205$. Agreement among four architecturally distinct recognizers instead reached an AUROC of $0.916$ ($[0.880,0.952]$). The two-of-four rule accepted $81.7\%$ of images at $88.8\%$ precision, the three-of-four rule $52.1\%$ at $98.5\%$. The same pattern held on CLEF-IP, UOB, and USPTO. This distinction is obscured on synthetic benchmarks, where re-rendered predictions naturally resemble their inputs. A substance filter removed $2{,}193$ false agreements on wildcards and R-group fragments, after which the three-of-four rule rejected all $68$ generic depictions. VERDICT was then applied to PMC Open Access, producing $6{,}146$ structure labels for $4{,}833$ molecules; chemist adjudication of $400$ released labels in two independent samples yielded precisions of $0.995$ for the three-of-four tier and $0.958$ for the two-of-four tier. VERDICT therefore enables validated labels for multimodal molecular databases linking structure images, machine-readable representations, and source-publication information. In SES AI's Molecular Universe platform, VERDICT further serves as an image-based interface for searching and retrieving molecular records.

发表机构

• SES AI Corporation（SES AI公司）

机构由 AI 辅助整理，请以论文原文为准。