RTHDet: Rotate Table Area and Head Detection in images

TL;DR

This paper introduces RTHDet, a novel method for detecting rotated table regions and localizing head-tail parts in images, addressing a gap in traditional horizontal table detection models, and demonstrates significant performance improvements with new datasets and metrics.

Contribution

The paper proposes RTHDet, a new detection framework with a rotated rectangle representation and angle loss, along with datasets and metrics for rotated table detection and localization.

Findings

RTHDet achieves AP50 (T<90) of 88.7%, significantly higher than baseline.

Introduces TRR360D dataset with semantic head-tail information.

Develops R360 AP metric for rotated region and part detection accuracy.

Abstract

Traditional models focus on horizontal table detection but struggle in rotating contexts, limiting progress in table recognition. This paper introduces a new task: detecting table regions and localizing head-tail parts in rotation scenarios. We propose corresponding datasets, evaluation metrics, and methods. Our novel method, 'Adaptively Bounded Rotation,' addresses dataset scarcity in detecting rotated tables and their head-tail parts. We produced 'TRR360D,' a dataset incorporating semantic information of table head and tail, based on 'ICDAR2019MTD.' A new metric, 'R360 AP,' measures precision in detecting rotated regions and localizing head-tail parts. Our baseline, the high-speed and accurate 'RTMDet-S,' is chosen after extensive review and testing. We introduce 'RTHDet,' enhancing the baseline with a 'r360' rotated rectangle angle representation and an 'Angle Loss' branch, improving head-tail localization. By applying transfer learning and adaptive boundary rotation augmentation, RTHDet's AP50 (T<90) improved from 23.7% to 88.7% compared to the baseline. This demonstrates RTHDet's effectiveness in detecting rotating table regions and accurately localizing head and tail parts.RTHDet is integrated into the widely-used open-source MMRotate toolkit: https://github.com/open-mmlab/mmrotate/tree/dev-1.x/projects/RR360.