现代信息科技2025，Vol.9Issue(22)：35-39,5.DOI:10.19850/j.cnki.2096-4706.2025.22.007
基于深度学习网络的AI拟声检测系统的设计
Design of AI Onomatopoeia Detection System Based on Deep Learning Network
摘要
Abstract
In the era of rapid development of information technology,AI onomatopoeia technology has brought changes to various industries,but it has also raised safety and ethical challenges,especially when it is difficult to distinguish between real human voices and AI-generated voices.The purpose of this paper is to design and implement an efficient AI onomatopoeia detection system to distinguish AI onomatopoeia from real human voices.The system uses a hybrid model method combining Recurrent Neural Network(RNN)and Variational Autoencoder(VAE)to learn the deep representation of sound features and capture the differences between AI onomatopoeia and real human voices.The sound is comprehensively analyzed from three dimensions of spectral characteristics,prosodic characteristics and sound quality.Then,the analysis of emotional speech synthesis technology is introduced to enhance the model's ability to recognize AI onomatopoeia.Experiments show that the detection system performs well on multiple public datasets,which proves its effectiveness and feasibility.关键词
AI拟声/语音合成/循环神经网络/变分自动编码器/情感语音合成/声音分析Key words
AI onomatopoeia/speech synthesis/Recurrent Neural Networks/variational autoencoder/affective speech synthesis/sound analysis分类
信息技术与安全科学引用本文复制引用
蓝家运,段晓霞,王丽颖..基于深度学习网络的AI拟声检测系统的设计[J].现代信息科技,2025,9(22):35-39,5.基金项目
大学生创新创业训练计划项目(XJ202313714344) （XJ202313714344）
