ACOUSTIC SCENE CLASSIFICATION USING DEEP NEURAL NETWORK AND FRAME-CONCATENATED ACOUSTIC FEATURE

This paper describes our contribution to the task of acoustic scene classification in the DCASE2016 (Detection and Classification of Acoustic Scenes and Events 2016) Challenge set by IEEE AASP. In this work, we applied the DNN-GMM (Deep Neural NetworkGaussian Mixture Model) to acoustic scene classification. We introduced high-dimensional features that are concatenated with acoustic features in temporally adjacent frames. As a result, it was confirmed that the classification accuracy of the DNN-GMM was improved by 5.0% in comparison with that of the GMM, which was used as the baseline classifier.