A Low-Compexity Deep Learning FrameworkFor Acoustic Scene Classification
In this paper, we presents a low-complexitydeep learning frameworks for acoustic scene classification(ASC). The proposed framework can be separated into threemain steps: Front-end spectrogram extraction, back-endclassification, and late fusion of predicted probabilities.First, we use Mel filter, Gammatone filter and ConstantQ Transfrom (CQT) to transform raw audio signal intospectrograms, where both frequency and temporal featuresare presented. Three spectrograms are then fed into threeindividual back-end convolutional neural networks (CNNs),classifying into ten urban scenes. Finally, a late fusion ofthree predicted probabilities obtained from three CNNs isconducted to achieve the final classification result. To reducethe complexity of our proposed CNN network, we applytwo model compression techniques: model restriction anddecomposed convolution. Our extensive experiments, whichare conducted on DCASE 2021 (IEEE AASP Challenge onDetection and Classification of Acoustic Scenes and Events)Task 1A development dataset, achieve a low-complexity CNNbased framework with 128 KB trainable parameters andthe best classification accuracy of 66.7%, improving DCASEbaseline by 19.0%.