An Acoustic Segment Model Based Segment Unit Selection Approach to Acoustic Scene Classification with Partial Utterances

Abstract

In this paper, we propose a sub-utterance unit selection framework to remove acoustic segments in audio recordings that carry little information for acoustic scene classification (ASC). Our approach is built upon a universal set of acoustic segment units covering the overall acoustic scene space. First, those units are modeled with acoustic segment models (ASMs) used to tokenize acoustic scene utterances into sequences of acoustic segment units. Next, paralleling the idea of stop words in information retrieval, stop ASMs are automatically detected. Finally, acoustic segments associated with the stop ASMs are blocked, because of their low indexing power in retrieval of most acoustic scenes. In contrast to building scene models with whole utterances, the ASM-removed sub-utterances, i.e., acoustic utterances without stop acoustic segments, are then used as inputs to the AlexNet-L back-end for final classification. On the DCASE 2018 dataset, scene classification accuracy increases from 68%, with whole utterances, to 72.1%, with segment selection. This represents a competitive accuracy without any data augmentation, and/or ensemble strategy. Moreover, our approach compares favourably to AlexNet-L with attention.

Relational Teacher Student Learning with Neural Label Embedding for Device Adaptation in Acoustic Scene Classification

“…In recent years, we have witnessed a great progress in the acoustic scene classification (ASC) task, as demonstrated by the high participation in the IEEE Detection and Classification of Acoustic Scenes and Events (DCASE) challenges [1,2,3]. Top ASC systems use deep neural networks (DNNs), and the main ingredient of their success is the application of deep convolutional neural networks (CNNs) [4,5,6,7,8,9]. Further boost in ASC performance is obtained with the introduction of advanced deep learning techniques, such as attention mechanism [10,11,12], mix-up [13,14], Generative Adversial Network (GAN) and Variational Auto Encoder (VAE) based data augmentation [15,16], and deep feature learning [17,18,19,20].…”