HTSTL: Head-and-Tail Search Network With Scale-Transfer Layer for Traffic Sign Text Detection

Xuan He, Runmin Wang, Xiumei Li, Xing Chen, Chengyu Guo, Lv Wen, Changxin Gao, Li Liu

Abstract

Traffic sign text detection is a vital component of autonomous driving systems, providing contextual information beyond simple sign recognition. However, environmental factors like low lighting and the geometric complexity of long, sparse text strings pose significant challenges. This paper presents HTSTL (Head-and-Tail Search Network), a robust detection pipeline featuring a Scale-Transfer Layer (STL) for efficient resolution recovery and a novel Head-and-Tail pixel classification strategy. Experimental results on the CTST-1600 dataset demonstrate that HTSTL achieves state-of-the-art performance (F-score of 0.82) while maintaining real-time processing speeds.

Introduction

Autonomous vehicles must interpret complex road environments in real-time. While object detection for standard traffic signs is well-established, "reading" the text on these signs remains difficult. Challenges include poor visibility at night, the extreme aspect ratios of Chinese traffic text, and the limited computational resources available on-board vehicles. Existing word-based detectors often fragment long text lines or fail in low-contrast scenarios.

Methodology

The HTSTL framework addresses these issues through a three-stage process: preprocessing, feature extraction via a modified Fully Convolutional Network (FCN), and structural post-processing.

Preprocessing and Scale-Transfer

To mitigate low-light failures, a preprocessing module applies Gamma correction based on the image's Root Mean Square (RMS) brightness. The core of the FCN architecture is the Scale-Transfer Layer (STL). Unlike traditional upsampling methods that introduce zero-padding or learnable parameters, STL rearranges channel-wise data into spatial dimensions. This approach restores resolution efficiently, reducing computational overhead.

Figure 1: The HTSTL pipeline, highlighting the preprocessing module and the integration of the Scale-Transfer Layer (STL) within the FCN.

Head-and-Tail Search Strategy

To handle long and sparse text, the network classifies pixels into three distinct categories: Head, Tail, and Middle. This structural classification allows the system to identify the boundaries of a text instance and link the internal components reliably, preventing the fragmentation common in standard bounding-box regression. By finding the "ends" of a sign and filling in the middle, the model treats detection as a structural search rather than a simple geometric exercise.

Figure 3: Visualization of the Head-and-Tail classification strategy. (a) Final bounding boxes. (b) Pixel-level classification where cyan represents 'Head', yellow represents 'Tail', and red represents 'Middle'.

Experimental Results

The model was evaluated on the CTST-1600 dataset, specifically curated for Chinese traffic signs, as well as standard benchmarks like ICDAR 2013 and MSRA-TD500.

  • Accuracy: HTSTL achieved an F-score of 0.82 on CTST-1600, significantly higher than the EAST (0.76) and TextSnake (0.64) models. On the ICDAR 2013 benchmark, it reached a precision of 0.92.
  • Efficiency: The system operates at approximately 5.28 FPS. While slightly slower than the EAST detector (6.33 FPS), it offers superior robustness in the complex scenarios for which it was designed.
  • Robustness: Qualitative results confirm the model's ability to detect text in near-total darkness where baseline models often fail.

Conclusion and Future Work

HTSTL provides a specialized solution for traffic sign text detection by combining efficient architectural components with a structurally-aware classification strategy. While highly effective for linear text in diverse lighting, future improvements will target curved text and overlapping instances. The authors also plan to integrate a text recognizer (OCR) directly into the pipeline to create an end-to-end reading system for autonomous navigation.

Summary updated

Abstract

Although promising results have been achieved in the area of traffic sign detection, little attention has been paid to text detection on traffic signs. In fact, in today's popular driver-less automobile industry, traffic sign text which brings abundant and valuable traffic information plays an important and indispensable role. In this work, we design an effective detector for traffic sign text, whose pipeline only consists of a preprocessing module to tackle with some complex situations, a Fully Convolutional Network (FCN) in which a Scale-transfer layer is proposed to speed up the network and a simple post-processing step. Extensive experiments on the Chinese traffic sign text dataset (CTST-1600), ICDAR 2013 and MSRA-TD500 show that the proposed method has achieved the state-of-the-art results, which proves the ability of our detector on both particularity and universality applications. We collect the Chinese text-based traffic sign dataset named CTST-1600, and it can be found at https://github.com/pummi823/test/blob/master/CTST-1600.
BibTeX
@article{he2019htstl,
  author = {Xuan He and Runmin Wang and Xiumei Li and Xing Chen and Chengyu Guo and Lv Wen and Changxin Gao and Li Liu},
  title = {{HTSTL: Head-and-Tail Search Network With Scale-Transfer Layer for Traffic Sign Text Detection}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2936540},
}
Figure