Evaluating Table Structure Recognition: A New Perspective

TL;DR

This paper introduces TEDS (IOU), a new metric for table structure recognition that uses bounding boxes to better evaluate alignment and robustness, addressing shortcomings of existing metrics.

Contribution

The paper proposes a novel TEDS (IOU) metric that improves evaluation of table structure recognition by focusing on bounding boxes and robustness.

Findings

TEDS (IOU) outperforms previous metrics in various examples.

The new metric better captures text and empty cell alignment issues.

Demonstrated robustness against common shortcomings of prior metrics.

Abstract

Existing metrics used to evaluate table structure recognition algorithms have shortcomings with regard to capturing text and empty cells alignment. In this paper, we build on prior work and propose a new metric - TEDS based IOU similarity (TEDS (IOU)) for table structure recognition which uses bounding boxes instead of text while simultaneously being robust against the above disadvantages. We demonstrate the effectiveness of our metric against previous metrics through various examples.