Robust color traffic sign recognition algorithm based on steerable pyramid transform and extreme learning machine

Abstract

In this work a new traffic sign recognition method by integrating color information based on Steerable Pyramid Transform (SPT) and Extreme Learning Machine (ELM) is introduced. The main contribution of this paper is to explore the usefulness of SPT at different color spaces in a traffic sign recognition framework. Each color traffic sign image is described by a subset of band filtered images containing steerable pyramid coefficients which characterize the traffic sign textures. Linear discriminant analysis is used to reduce the data dimensionality to generate relevant features. These reduced features are used as the input to ELM classifier to analytically learn an optimal model. This work can be a stepping stone for further research in this direction. Experiments on German Traffic Sign Recognition Benchmark (GTSRB) and Belgium Traffic Sign Classification Benchmark (BTSCB) databases demonstrate that our proposed method color SPT-ELM can serve as an effective and reasonable feature extraction tool, and achieve good and fast recognition accuracy. In addition, comparisons against some state-of-the-art methods prove the effectiveness and the superiority of the proposed approach for color traffic sign recognition and it is more adequate for real-time application.