Improving the Robustness and Clinical Applicability of Automatic Respiratory Sound Classification Using Deep Learning–Based Audio Enhancement: Algorithm Development and Validation｜BIIC Lab - NTHU

Spoken Dialogs

Speech and Language

Improving the Robustness and Clinical Applicability of Automatic Respiratory Sound Classification Using Deep Learning–Based Audio Enhancement: Algorithm Development and Validation

PDF JMIR AI

Abstract

000

Figures

Figure 1. An overview of our proposed noise-robust respiratory sound classification system with audio enhancement. CNN: convolutional neural network; CNN14: 14-layer CNN; conformer: convolution-augmented transformer; ISTFT: inverse short-time Fourier transform; STFT: short-time Fourier transform; TS: 2 stage.

Keywords

respiratory sound ｜ lung soun ｜ audio enhancement ｜ noise robustness ｜ clinical applicability ｜ artificial intelligence ｜ AI

Authors

Publication Date

-0001/11/30

Journal

JMIR AI

DOI

doi: 10.2196/67239

RESEARCH

Related Research