CROHME Contextual Dataset

This dataset supports CROHME2019 [1] and CROHME2023 [2] Testset with contextual information. It provides aligned LaTeX expressions extracted from source documents, along with surrounding textual context, to facilitate formula recognition experiments. The paper is publised in ICDAR 2025 M3RD Workshop [3]. The dataset contains two main components: the original source files and processed JSON data aligned with CROHME test sets.本数据集为带有上下文信息的CROHME2019[1]与CROHME2023[2]测试集提供支撑。其提供了从源文档中提取的对齐后LaTeX表达式，以及周边文本上下文，以助力公式识别相关实验。本论文已发表于ICDAR 2025 M3RD研讨会[3]。 该数据集包含两大核心组成部分：原始源文件，以及与CROHME测试集对齐的经处理JSON数据。