arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2007.15283eess.AScs.LGcs.SD
收藏

用于深度说话人嵌入的特征提取器的比较再评估

A Comparative Re-Assessment of Feature Extractors for Deep Speaker Embeddings

University of Eastern Finland(东芬兰大学);CNRS(法国国家科学研究中心);Inria(法国国家信息与自动化研究所)

Xuechen Liu, Md Sahidullah, Tomi Kinnunen

摘要:

现代自动说话人验证在很大程度上依赖于在梅尔频率倒谱系数(MFCC)特征上训练的深度神经网络(DNN)。虽然有基于相位,韵律和长期时间操作的替代特征提取方法,但它们尚未被基于DNN的方法广泛研究。我们的目标是通过对VoxCeleb和SITW数据集上的14个特征提取器进行广泛的重新评估来填补这一空白。我们的研究结果表明,配备频谱质心,群延迟函数和综合噪声抑制等技术的功能为深层说话人嵌入提取提供了有前途的替代方案。实验结果表明,等错误率(EER)相对于基线降低了16.3%(VoxCeleb)和25.1%(SITW)。

英文摘要:

Modern automatic speaker verification relies largely on deep neural networks (DNNs) trained on mel-frequency cepstral coefficient (MFCC) features. While there are alternative feature extraction methods based on phase, prosody and long-term temporal operations, they have not been extensively studied with DNN-based methods. We aim to fill this gap by providing extensive re-assessment of 14 feature extractors on VoxCeleb and SITW datasets. Our findings reveal that features equipped with techniques such as spectral centroids, group delay function, and integrated noise suppression provide promising alternatives to MFCCs for deep speaker embeddings extraction. Experimental results demonstrate up to 16.3\% (VoxCeleb) and 25.1\% (SITW) relative decrease in equal error rate (EER) to the baseline.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑