Results of our traffic sign recognition system (detection + classification over 62 classes) shown on the Belgian traffic sign dataset.
Evaluation is done frame by frame without any temporal consistency or geometric prior.
Our detector is one of the winning approaches of the German Traffic Sign Detection Benchmark (GTSDB), “visics” entry at:
https://benchmark.ini.rub.de/?section=gtsdb&subsection=results
Details in:
“Traffic Sign Recognition – How far are we from the solution?”
M. Mathias, R. Timofte, R. Benenson, and L. Van Gool.
IJCNN 2013
https://markusmathias.de/publications/2013_ijcnn_traffic_signs.pdf
Related work and more details at:
https://markusmathias.de/publications/
and
https://www.vision.ee.ethz.ch/~timofter/
Components of our system have been previously used successfully for pedestrian detection and face recognition.
Post time: Jun-19-2017