Overlapping acoustic event detection via perceptually inspired the holistic-based representation method

Hyeonsik Choi, Keunsang Lee, Minseok Keum, David Han, Hanseok Ko

Research output: Contribution to conferencePaperpeer-review

Abstract

A novel dictionary learning approach that utilizes Mel-scale frequency warping in detecting overlapped acoustic events is proposed. The study explored several dictionary learning schemes for improved performance of overlapping acoustic event detection. The structure of NMF for calculating gains of each event was utilized for including in overlapped signal for its low computational load. In this paper, we propose a method of frequency warping for better sound representation, and apply dictionary learning by a holistic-based representation, namely nonnegative K-SVD (NK-SVD) in order to resolve a basis sharing problem raised by part-based representations. By using Mel-scale in a dictionary learning, we show that the information carried by low frequency components more than high frequency components and dealt with a low complexity. Also, the proposed holistic-based representation method avoids the permutation problem between another acoustic events. Based on these benefits, we confirm that the proposed method of Mel-scale with NK-SVD delivered significantly better results than the conventional methods.

Original languageEnglish
Publication statusPublished - 2020
Event149th Audio Engineering Society Convention 2020, AES 2020 - Virtual, Online
Duration: 2020 Oct 272020 Oct 30

Conference

Conference149th Audio Engineering Society Convention 2020, AES 2020
CityVirtual, Online
Period20/10/2720/10/30

ASJC Scopus subject areas

  • Modelling and Simulation
  • Acoustics and Ultrasonics

Fingerprint Dive into the research topics of 'Overlapping acoustic event detection via perceptually inspired the holistic-based representation method'. Together they form a unique fingerprint.

Cite this