Context-Aware Chart Element Detection

TL;DR

This paper introduces CACHED, a novel context-aware method for detecting chart elements that significantly improves accuracy by integrating local and global context information, and achieves state-of-the-art results in chart element detection and bar plot recognition.

Contribution

The paper proposes CACHED, a new context-aware detection framework that incorporates visual and positional context, and refines chart element categories for better generalization.

Findings

Achieves state-of-the-art performance in chart element detection.

Extends effectively to bar plot detection with top results.

Highlights the importance of context in structured data visualization detection.

Abstract

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.