Road Sign Detection in Images: A Case Study

TLDR

Road sign identification is crucial for vehicle safety and is typically addressed in detection, recognition, and tracking stages evaluated together. This study concentrates on improving road sign detection, the first stage of the overall process. The authors compare three state‑of‑the‑art detection algorithms—Contour Fitting, Radial Symmetry Transform, and pair‑wise voting—on a benchmark of 847 urban images containing 251 signs, using color, edge, and geometric models. The comparative analysis highlights each algorithm’s strengths and weaknesses, suggesting directions for future research.

Abstract

Road sign identification in images is an important issue, in particular for vehicle safety applications. It is usually tackled in three stages: detection, recognition and tracking, and evaluated as a whole. To progress towards better algorithms, we focus in this paper on the first stage of the process, namely road sign detection. More specifically, we compare, on the same ground-truth image database, results obtained by three algorithms that sample different state-of-the-art approaches. The three tested algorithms: Contour Fitting, Radial Symmetry Transform, and pair-wise voting scheme, all use color and edge information and are based on geometrical models of road signs. The test dataset is made of 847 images 960×1080 of complex urban scenes (available at www.itowns.fr/benchmarking.html). They feature 251 road signs of different shapes (circular, rectangular, triangular), sizes and types. The pros and cons of the three algorithms are discussed, allowing to draw new research perspectives.