融合注意力机制的ResNeXt语音欺骗检测模型
SPEECH ANTI-SPOOFING MODEL BASED ON RESNEXT WITH ATTENTION
张旺 1杨乘 2罗娅娅1
作者信息
- 1. 贵州省教育厅汽车电子技术特色重点实验室贵州师范大学物理与电子科学学院 贵州贵阳 550025
- 2. 贵州省教育厅汽车电子技术特色重点实验室贵州师范大学物理与电子科学学院 贵州贵阳 550025;贵州省射电天文数据处理重点实验室贵州师范大学物理与电子科学学院 贵州贵阳 550025
- 折叠
摘要
针对残差神经网络在语音欺骗检测中存在超参数过多且对于高频特征显著性突出不够的问题,提出一种融合注意力机制的ResNeXt-Attention网络(RA-Net).RA-Net采用残差结合分组卷积的方式,用一组小卷积核代替大卷积核,且采用MFM(Max Feature Map)作为新的拼接方法.加入的注意力机制通过学习原始特征的信息,减少了对边缘信息的关注.在ASVspoof2019数据集上实验表明,RA-Net相比基准线高斯混合模型(GMM)的等错误率(EER)降低了 4.72百分点和6.23百分点,与残差网络(Residal Neural Network,ResNet)相比EER降低了 0.69百分点和0.89百分点,证明了该模型的有效性.
Abstract
Aimed at the problem that residual neural network has too many hyperparameters in speech deception detection,and the high-frequency features are not prominent enough,a ResNeXt-Attention network(RA-Net)fused with attention mechanism is proposed.RA-Net used residuals combined with grouped convolution,replaced large convolution kernels with a set of small convolution kernels,and used MFM(max feature map)as a new splicing method.The added attention mechanism reduced the attention to edge information by learning the original feature information.Experiments on the ASVspoof2019 data set show that compared with the baseline Gaussian mixture model(GMM),the equal error rate(EER)of RA-Net is reduced by 4.72 percentage points and 6.23 percentage points.And the EER is reduced by 0.69 percentage points and 0.89 percentage points compared with the residual network(ResNet).The validity of the model is confirmed.
关键词
语音欺骗检测/ResNeXt/MFM/注意力机制/RA-NetKey words
Speech anti-spoofing/ResNeXt/MFM/Attention mechanism/RA-Net引用本文复制引用
基金项目
国家自然科学基金项目(62062025)
国家自然科学基金项目(61662010)
贵州省科学技术基金重点项目(黔科合基础[2019]1432)
出版年
2024
国家科技期刊平台
NSTL
万方数据