The German Traffic Sign Benchmark

http://benchmark.ini.rub.de/?section=gtsrb&subsection=about

Overview

The german traffic sign benchmark dataset consists of over 50 000 image samples presenting more than 40 types of traffic signs (thus, 40 classes), alongside semi-automatic annotations. Each image contains exactly one traffic sign, with a border of about 5 pixels around to allow edge-based detection.

The actual traffic sign is not necessarily centered within the image.This is true for images that were close to the image border in the full camera image.

Associated Paper or Article

More information can be found by reading Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognition.

Annotations

Download

The dataset can be downloaded here.

Additional Python code for reading all images can be accessed here.

Model

A trained model can be found here. Please also read the README file.

The authors note that the code is platform-independent, however, it has only been tested Visual Studio. So there might be a couple of issues left where GCC is more strict than Visual Studio. We appreciate any comments, patches and bug reports. The code uses CMake, an open-source, cross-platform build system which allows you to generate project files/makefiles for your preferred compiler toolchain.

Benchmarks

The official benchmark can be consulted here.

Associated Challenges

This dataset has been associated with the IJCNN 2011 competition (1st stage).

The competition results can be consulted here.

License

No license found.