![](https://img-blog.csdnimg.cn/20201014180756928.png?x-oss-process=image/resize,m_fixed,h_64,w_64)
语音验证
文章平均质量分 94
FallenDarkStar
怀着狂妄理想不自量力地前进着
展开
-
【论文学习】ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation
目前的说话人验证技术依赖于神经网络来提取说话人的表征。成功的x-vector架构是一个时延神经网络(TDNN),它应用统计池化将变长语句投影到表征嵌入的定长说话人中。本文基于人脸验证和计算机视觉相关领域的最新趋势,对该架构提出了多种增强。首先,将初始帧层重构为具有有效跳跃连接的1维Res2Net模块;与SE-ResNet类似,在这些模块中引入了挤压和激发块,以显式地对通道相互依赖性进行建模。SE块根据记录的全局特性重新缩放通道,扩展帧层的时间上下文。原创 2023-05-09 14:27:36 · 2094 阅读 · 1 评论 -
【论文学习】《Generalized End-to-End Loss for Speaker Verification》
《Generalized End-to-End Loss for Speaker Verification》论文学习文章目录《Generalized End-to-End Loss for Speaker Verification》论文学习 摘要 1 介绍 1.1 背景 1.2 基于元组的端到端损失 1.3 概述 2 广义的端到端模型 2.1 训练方法 &原创 2021-11-26 12:42:18 · 2732 阅读 · 0 评论 -
【论文学习】《A Overview of Spoof Speech Detection for Automatic Speaker Verification》
《A Overview of Spoof Speech Detection for Automatic Speaker Verification》论文学习文章目录《A Overview of Spoof Speech Detection for Automatic Speaker Verification》论文学习 摘要 1 介绍 2 ASV 系统:欺骗攻击 2.1 双胞胎 2.2 模仿 2.原创 2021-10-16 13:47:11 · 3244 阅读 · 0 评论 -
【论文学习】《Replay attack detection with complementary high-resolution information using end-to-end DNN 》
《Replay attack detection with complementary high-resolution information using end-to-end DNN for the ASVspoof 2019 Challenge》论文学习文章目录《Replay attack detection with complementary high-resolution information using end-to-end DNN for the ASVspoof 2019 Challen原创 2021-05-22 17:42:33 · 1336 阅读 · 4 评论 -
【论文学习】ASVspoof 2015 the First Automatic Speaker Verification Spoofing and Countermeasures Challenge
Lout=⌊Lin+2×padding−dilation×(kernel_size−1)−1stride+1⌋L_{out} = \left\lfloor \frac{L_{in} + 2 \times \text{padding} - \text{dilation} \times (\text{kernel\_size} - 1) - 1}{\text{stride}} + 1\right\rfloorLout=⌊strideLin+2×padding−dila原创 2021-05-20 18:59:10 · 2021 阅读 · 12 评论