Towards a robust vision-based obstacle perception with classifier fusion in cybercars

Abstract

Several single classifiers have been proposed to recognize objects in images. Since this approach has restrictions when applied in certain situations, one has suggested some methods to combine the outcomes of classifiers in order to increase overall classification accuracy. In this sense, we propose an effective method for a frame-by-frame classification task, in order to obtain a trade-off between false alarm decrease and true positive detection rate increase. The strategy relies on the use of a Class Set Reduction method, using a Mamdani fuzzy system, and it is applied to recognize pedestrians and vehicles in typical cybercar scenarios. The proposed system brings twofold contributions: i) overperformance with respect to the component classifiers and ii) expansibility to include other types of classifiers and object classes. The final results have shown the effectiveness of the system1. © Springer-Verlag Berlin Heidelberg 2007.