A Novel Approach to Traffic Sign Recognition Based on the Optimization Under Transformations

Abstract

An automatic traffic sign recognition system can provide drivers with extremely important information for safe and successful driving. This paper proposes an automatic approach for traffic sign recognition in uncontrolled environments. The main idea of this method is to transform the patterns of the traffic signs into an optimal state defined by a function on the transformations. Under the optimal state, the undesired transformation effects on the traffic sign patterns are removed, and this can help improving the precision of traffic sign recognition. A coarse to fine traffic sign recognition system based on template matching is designed in the consideration of both efficiency and precision. Experimental results showed that the proposed approach performed well for traffic sign recognition.