Traffic sign recognition based on stable SURFs
基于高稳定SURF特征的交通标志识别

This paper proposed a new method to recognize traffic signs based on SURFs with high stability. A set of SURFs, which was called bag of SURFs, was constructed using SURFs from multiple video frames that contain a traffic sign, and then matched with a set of SURFs from a standard traffic sign image. The recognition result was determined by the highest score of a weighted scoring scheme. Carried out a experiments based on a public dataset under three given conditions, and also performed comparative analysis with two recent methods. The results demonstrate that the method is an efficent traffic sign recognition method and yields a significant superiority under variable lighting or partial condition.