Self-subtraction network for end to end noise robust classification

Donghyeon Kim, David K. Han, Hanseok Ko

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Acoustic event classification in surveillance applications typically employs deep learning-based end-to-end learning methods. In real environments, their performance degrades significantly due to noise. While various approaches have been proposed to overcome the noise problem, most of these methodologies rely on supervised learning-based feature representation. Supervised learning system, however, requires a pair of noise free and noisy audio streams. Acquisition of ground truth and noisy acoustic event data requires significant efforts to adequately capture the varieties of noise types for training. This paper proposes a novel supervised learning method for noise robust acoustic event classification in an end-to-end framework named Self Subtraction Network (SSN). SSN extracts noise features from an input audio spectrogram and removes them from the input using LSTMs and an auto-encoder. Our method applied to Urbansound8k dataset with 8 noise types at four different levels demonstrates improved performances compared to the state of the art methods.

Original languageEnglish
Title of host publication2019 16th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728109909
DOIs
Publication statusPublished - 2019 Sep
Event16th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2019 - Taipei, Taiwan, Province of China
Duration: 2019 Sep 182019 Sep 21

Publication series

Name2019 16th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2019

Conference

Conference16th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2019
CountryTaiwan, Province of China
CityTaipei
Period19/9/1819/9/21

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Vision and Pattern Recognition
  • Signal Processing

Fingerprint Dive into the research topics of 'Self-subtraction network for end to end noise robust classification'. Together they form a unique fingerprint.

  • Cite this

    Kim, D., Han, D. K., & Ko, H. (2019). Self-subtraction network for end to end noise robust classification. In 2019 16th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2019 [8909821] (2019 16th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/AVSS.2019.8909821