Design of voiceprint recognition algorithm based on triplet-loss-optimized deep learning
Zhendong Xu, Feng Gao, Yuebo Jia, Ziyu Gu, Yande Liu, Jun Xu · 2024
Voiceprint recognition is a technology that verifies or identifies individuals by analysing their speech signals. Traditional voiceprint recognition methods have issues like inaccurate feature extraction and weak robustness. To enhance the performance of voiceprint recognition, this paper adopts the triplet loss and its optimization. By employing the optimized triplet loss, we can effectively improve the accuracy and robustness of voiceprint recognition.