摘要
随着科技的不断发展,生成对抗网络在生成图像领域取得的重要进展超过了诸多经典方法。文章对基于GAN生成图像特征检测技术研究进行了综述,首先简述了生成对抗网络的基本原理及深度伪造制品中的应用场景,其次对目前GAN特征检测技术进行了分类,最后探讨了深度伪造技术传播带来的社会风险,并分析了现有检测技术的不足,可为检测技术未来的发展和应用提供参考。
With the continuous development of technology,generative adversarial networks have made significant progress in the field of image generation,surpassing many classic methods.The article provides a review of research on image feature detection technology based on GAN.Firstly,the basic principles of generative adversarial networks and their application scenarios in deep forgery products are briefly described.Secondly,the current GAN feature detection technology is classified.Finally,the social risks brought by the dissemination of deep forgery technology are discussed,and the shortcomings of existing detection technologies are analyzed,which can provide reference for the future development and application of detection technology.
作者
黄勐
苏金善
HUANG Meng;SU Jinshan(School of Electronic and Engineering,Yili Normal University,Yining,Xinjiang 835000,China)
出处
《计算机应用文摘》
2024年第3期90-92,共3页
Chinese Journal of Computer Application
基金
新疆自然科学联合基金(2016D01C382)
新疆维吾尔自治区研究生创新项目(XJ2023G258)。
关键词
生成对抗网络
虚假图像
特征检测
深度学习
generative adversarial network
fake image
feature detection
deep learning
