A voiceprint anti-counterfeiting recognition method and system based on voiceprint feature time sequence track

By analyzing the temporal trajectory of voiceprint features, feature vectors are directly extracted using a pre-trained model, and similarity and smoothness scores are calculated. This solves the problem that existing voiceprint recognition technologies cannot defend against splicing attacks, and achieves efficient, secure, and interpretable anti-spoofing voiceprint recognition.

CN122417033APending Publication Date: 2026-07-17RONGZHITONG TECH BEIJING
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
RONGZHITONG TECH BEIJING
Filing Date
2026-04-16
Publication Date
2026-07-17

Smart Images

  • Figure CN122417033A_ABST
    Figure CN122417033A_ABST
Patent Text Reader

Abstract

本发明公开了一种基于声纹特征时序轨迹的防伪造声纹识别方法及系统，属于生物特征识别与信息安全技术领域。所述方法包括：对输入音频进行预处理与重叠滑动切片，提取按时间排序的音频子片段序列；将子片段输入声纹特征提取模型，得到高维声纹特征向量时序序列，形成声纹特征时序轨迹；计算特征向量与目标用户注册特征的相似度得到身份得分，进行主路身份验证；分析特征时序轨迹的一致性，量化轨迹平滑度得到平滑度得分，进行辅路防伪验证；通过身份得分与平滑度得分的联合逻辑判决，判定语音是否为真实语音。本发明无需额外训练反欺诈模型，可无缝集成到现有声纹识别系统，适用于金融支付、门禁考勤等多场景的声纹防伪识别。
Need to check novelty before this filing date? Find Prior Art