Explicit Relational Reasoning Network for Scene Text Detection

TL;DR

ERRNet introduces an end-to-end scene text detection method that models component relationships explicitly, eliminating post-processing and achieving state-of-the-art accuracy with high efficiency.

Contribution

The paper proposes ERRNet, a novel relational reasoning network that treats text components as objects in a tracking framework, removing the need for post-processing in CC-based text detection.

Findings

Achieves state-of-the-art accuracy on benchmarks.

Eliminates post-processing in scene text detection.

Maintains high inference speed.

Abstract

Connected component (CC) is a proper text shape representation that aligns with human reading intuition. However, CC-based text detection methods have recently faced a developmental bottleneck that their time-consuming post-processing is difficult to eliminate. To address this issue, we introduce an explicit relational reasoning network (ERRNet) to elegantly model the component relationships without post-processing. Concretely, we first represent each text instance as multiple ordered text components, and then treat these components as objects in sequential movement. In this way, scene text detection can be innovatively viewed as a tracking problem. From this perspective, we design an end-to-end tracking decoder to achieve a CC-based method dispensing with post-processing entirely. Additionally, we observe that there is an inconsistency between classification confidence and localization quality, so we propose a Polygon Monte-Carlo method to quickly and accurately evaluate the localization quality. Based on this, we introduce a position-supervised classification loss to guide the task-aligned learning of ERRNet. Experiments on challenging benchmarks demonstrate the effectiveness of our ERRNet. It consistently achieves state-of-the-art accuracy while holding highly competitive inference speed.