Deep Neural Network Based Supervised Speech Segregation Generalizes to Novel Noises through Large-scale Training

Abstract

Deep neural network (DNN) based supervised speech segregation has been successful in improving human speech intelligibility in noise, especially when DNN is trained and tested on the same noise type. A simple and effective way for improving generalization is to train with multiple noises. This letter demonstrates that by training with a large number of different noises, the objective intelligibility results of DNN based supervised speech segregation on novel noises can match or even outperform those on trained noises. This demonstration has an important implication that improving human speech intelligibility in unknown noisy environments is potentially achievable.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2015
Accession Number
AD1001124

Entities

People

  • DeLiang Wang
  • Jitong Chen
  • Yuxuan Wang

Organizations

  • Ohio State University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Background Noise
  • Computer Science
  • Computers
  • Detectors
  • Engineering
  • Estimators
  • False Alarms
  • Frequency
  • Frequency Bands
  • Human Factors Engineering
  • Intelligibility
  • Machine Learning
  • Neural Networks
  • Supervised Machine Learning
  • Test Sets
  • Training
  • Warning Systems

Fields of Study

  • Computer science

Readers

  • Neural Network Machine Learning.
  • Speech Processing/Speech Recognition.

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks