Acoustic vehicle classification by fusing with semantic annotation

Current research on acoustic vehicle classification has been generally aimed at utilizing various feature extraction methods and pattern recognition techniques. Previous research in gait biometrics has shown that domain knowledge or semantic enrichment can assist in improving the classification accuracy. In this paper, we address the problem of semantic enrichment by learning the semantic attributes from the training set, and then formalize the domain knowledge by using ontologies. We first consider a simple data ontology, and discuss how to use it for classification. Next we propose a scheme, which uses a semantic attribute to mediate information fusion for acoustic vehicle classification. To assess the proposed approaches, experiments are carried out based on a data set containing acoustic signals from five types of vehicles. Results indicate that whether the above semantic enrichment can lead to improvement depends on the accuracy of semantic annotation. Among the two enrichment schemes, semantically mediated information fusion achieves less significant improvement, but is insensitive to the annotation error.