作为图表数据提取的前提，准确检测图表基本元素是必不可少且强制的。与一般图像领域中的目标检测不同，图表元素检测严重依赖上下文信息，因为图表是高度结构化的数据可视化格式。为了解决这一问题，我们提出了一种新方法CACHED（Context-Aware Chart Element Detection，上下文感知图表元素检测），通过将包含视觉上下文增强和位置上下文编码的局部-全局上下文融合模块与Cascade R-CNN框架相集成来实现。为了提高我们方法的泛化能力以实现更广泛的适用性，我们改进了现有的图表元素分类，并标准化了18类图表基本元素（不包括绘图元素）。我们的CACHED方法结合更新后的图表元素类别，在实验中取得了最先进的性能，凸显了上下文在图表元素检测中的重要性。将我们的方法扩展到条形图检测任务时，我们在PMC测试数据集上获得了最佳结果。
英文摘要：
As a prerequisite of chart data extraction, the accurate detection of chart basic elements is essential and mandatory. In contrast to object detection in the general image domain, chart element detection relies heavily on context information as charts are highly structured data visualization formats. To address this, we propose a novel method CACHED, which stands for Context-Aware Chart Element Detection, by integrating a local-global context fusion module consisting of visual context enhancement and positional context encoding with the Cascade R-CNN framework. To improve the generalization of our method for broader applicability, we refine the existing chart element categorization and standardized 18 classes for chart basic elements, excluding plot elements. Our CACHED method, with the updated category of chart elements, achieves state-of-the-art performance in our experiments, underscoring the importance of context in chart element detection. Extending our method to the bar plot detection task, we obtain the best result on the PMC test dataset.
Towards an efficient framework for Data Extraction from Chart Images
South China University of Technology（华南理工大学）；IntSig Information Co.,Ltd.（合合信息科技股份有限公司）；Lenovo Research（联想研究院）；Guangdong Artificial Intelligence and Digital Economy Laboratory (Pazhou Lab)（广东人工智能与数字经济实验室（琶洲实验室））
