统计,多尺度和基于注意力的Wav2Vec-2语音嵌入层聚合用于帕金森病检测
Ondrej Klempir1, Juliana Grand Mullerova2, Radim Krupicka1
1Department of Biomedical Informatics, Faculty of Biomedical Engineering, Czech Technical University in Prague, Sitna Square 3105, Kladno, Czech Republic.
Computers in biology and medicine
|December 6, 2025
概括
平均汇集有效地汇总了用于帕金森病 (PD) 检测的语音模型嵌入. 这种简单的方法优于复杂的技术,提供了使用 wav2vec 2.0 嵌入式的强大而准确的PD 语音选.
科学领域:
- 语音处理 语音处理
- 机器学习在医疗保健中的应用
- 生物医学信号分析分析
背景情况:
- 像 wav2vec 2.0 这样的自我监督模型为临床应用产生了有价值的语音嵌入.
- 将框架级嵌入到发言描述器中的聚合对于像帕金森病 (PD) 查等任务至关重要.
- 使用这些嵌入式的PD检测的最佳聚合策略仍未得到充分探索.
研究的目的:
- 为了比较PD检测中 wav2vec 2.0 语音嵌入的各种聚合方法.
- 评估不同模型变体,层深度和聚合函数的影响.
- 为二进制分类 (PD与健康对照) 创建固定长度发言描述符的最有效策略.
主要方法:
- 在多个层深度中比较了三种wav2vec 2.0变体 (基础,微调).
- 评估了十个统计聚合函数 (平均值,中位数,量值) 和两个高级方案 (注意力,多尺度聚合).
- 使用了MDVR-KCL读音语语料库 (16 PD, 21 HC) 并应用监督特征选择 (ANOVA F值).
主要成果:
- 简单的统计聚合,特别是平均值聚合,始终优于复杂的方法.
- 早期的模型表示 (变压器前,第1变压器块) 通常是最有信息的.
- 基于wav2vec的嵌入式超越了传统的声学基线,特征选择提高了性能 (例如,0.87平衡精度,0.92精度).
结论:
- 均值聚合是一种强大而有效的策略,用于PD检测中wave2vec 2.0嵌入的时间聚合.
- 简单的聚合方法通常是足够的,并且比复杂的方法更可靠.
- 当适当聚合时, wav2vec 2.0 嵌入式为非侵入性 PD 查提供了一个有希望的途径.
相关概念视频
Parkinson's Disease: Overview
1.7K
Neurodegenerative disorders are progressive diseases that cause irreversible damage and loss to neurons in specific brain areas. Examples of these disorders include Parkinson's disease, Alzheimer's disease, Multiple Sclerosis (MS), and Amyotrophic Lateral Sclerosis (ALS). These disorders share characteristics such as proteinopathies, selective neuronal vulnerability, and a complex interplay between genetic and environmental factors. The primary therapeutic goal for these conditions is...
1.7K
Parkinson's Disease: Treatment
945
Neurodegenerative disorders, such as Parkinson's Disease (PD), involve the gradual and irreversible destruction of neurons in particular brain areas. These disorders exhibit standard features like proteinopathies, selective vulnerability of some neurons, and an interaction of intrinsic properties, genetics, and environmental influences in neural injury.
Parkinson's Disease is primarily a result of the loss of dopaminergic neurons in the substantia nigra pars compacta. The cornerstone of...
Parkinson's Disease is primarily a result of the loss of dopaminergic neurons in the substantia nigra pars compacta. The cornerstone of...
945


