Download PDFOpen PDF in browser
EN
The title and the abstract of this preprint are also available
in English

基于卷积神经网络的单一特征及多特征融合阿尔茨海默症音频分类方法

EasyChair Preprint 6747

9 pagesDate: October 3, 2021

Abstract

阿尔茨海默症是一种渐进式脑部疾病,随着时间的推移而恶化,因此对阿尔茨海默症进行早期筛查具有非常重要的意义。此前对阿尔茨海默症的识别依赖于各类声学特征和语音转录,而本文仅利用说话人的音频特征,提出了基于卷积神经网络多特征融合的阿尔茨海默症音频识别方法,并结合模型间集成学习方法,利用多次随机采样、语音端点检测等语音处理策略,实现了对阿尔兹海默症患者、轻度认知障碍患者及正常人三类人群音频的有效区分。本文提出的模型在长音频赛道和短音频赛道分别取得了84.87%(排名第四)和83.78%(排名第三)的准确率,此外,本文还探究了在不同网络结构及训练方法上的阿尔兹海默症识别表现。本文的源代码可以在https://github.com/lzl32947/NCMMSC2021_AD_Competition中访问。

Keyphrases: Alzheimer detection, audio classification, machine learning, multi-feature fusion, speech processing

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
@booklet{EasyChair:6747,
  author    = {Zhilin Liu and Yanyu Yang and Zhao Yang and Kun Zhao and Wei Xi},
  title     = {Single-Feature and Multi-Feature Fusion Audio Classification for Alzheimer's Disease Based on Convolutional Neural Network},
  howpublished = {EasyChair Preprint 6747},
  year      = {EasyChair, 2021}}
Download PDFOpen PDF in browser