Summary
International Workshop on Smart Info-Media Systems in Asia
2023
Session Number:RS2
Session:
Number:RS2-6
F-Cutmix: Investigation of data augmentationusing flexible Cutmix for speech enhancement
Reito Kasuga, Yosuke Sugiura, Tetsuya Shimamura,
pp.57-61
Publication Date:2023/8/31
Online ISSN:2188-5079
DOI:10.34385/proc.77.RS2-6
PDF download (2.6MB)
Summary:
Recently, deep-learning-based techniques have become benchmark models for speech enhancement, and their performance has been updated. Nevertheless, there is a problem of low versatility for various noise environments due to the lack of large-scale speech datasets. In this paper, we attempt to solve the data size problem and improve speech enhancement performance by introducing data augmentation methods "F-Cutmix" to speech enhancement networks.