DeepFakes分析与检测研究
首发时间:2021-01-12
摘要:AI 视频换脸技术滥用的安全问题日益严峻。现有换脸应用在 Deepfakes 基础上开发,研究 Deepfakes检测技术对整个领域具有重要意义。为满足 DeepFakes 应用安全性需求,本文系统分析总结现有检测技术;首次从取证角度研究,提出可行检测技术并进行初步验证;首次从生成式对抗网络特点切入提出可行检测方法。
关键词: DeepFakes; 图像篡改检测; 生成式对抗网络; 卷积神经网络
For information in English, please click here
Research of DeepFakes Analysis and Detection Methods
Abstract:The security problem of AI video face replacement technology is becoming more and more serious.The existing face-changing application is developed on the basis of Deepfakes, and the research on Deepfakes detection technology is of great significance to the whole field. In order to meet the security requirements of DeepFakes application, this paper firstly analyzes and summarizes the existing detection technology. For the first time, from the perspective of forensic research, it proposes feasible detection technology and conducts preliminary verification. For the first time, it proposes a feasible detection method from the characteristics of the generated confrontation network.
Keywords: DeepFakes Image tamper detection GAN CNN
引用
No.****
同行评议
勘误表
DeepFakes分析与检测研究
评论
全部评论0/1000