DEEPFAKE DETECTION USING LSTM & RESNEXT-50 1

M. Sm
{"title":"DEEPFAKE DETECTION USING LSTM & RESNEXT-50 1","authors":"M. Sm","doi":"10.55041/ijsrem34528","DOIUrl":null,"url":null,"abstract":"As computing power increases, \"deep fakes,\" or films that seem like they were created by a real person, are becoming feasible thanks to deep learning algorithms. Political instability, terrorism, revenge porn, or extortion can be the motivation for these lifelike face swapped deep fakes. This research presents a novel deep learning system that can recognize the difference between real and AI-generated films.We have developed a method that can identify deep fake replacements and reenactments automatically. We're using AI to combat AI. Our system's Res-Next Convolution neural network retrieves properties at the frame level. A recurrent neural network (RNN) trained using long short-term memory (LSTM) characteristics can distinguish between deep films and regular movies. On big, balanced, and mixed datasets, our technique is tested by Face-Forensic++ [1], Deepfake Detection Challenge [2], and Celeb-DF [3]. The quality of real-time data is enhanced by this. We prove that our technology consistently outperforms the competitors. Computer vision, Res-Next Convolution neural network, RNN, and LSTM are some of the index phrases.","PeriodicalId":13661,"journal":{"name":"INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.55041/ijsrem34528","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

As computing power increases, "deep fakes," or films that seem like they were created by a real person, are becoming feasible thanks to deep learning algorithms. Political instability, terrorism, revenge porn, or extortion can be the motivation for these lifelike face swapped deep fakes. This research presents a novel deep learning system that can recognize the difference between real and AI-generated films.We have developed a method that can identify deep fake replacements and reenactments automatically. We're using AI to combat AI. Our system's Res-Next Convolution neural network retrieves properties at the frame level. A recurrent neural network (RNN) trained using long short-term memory (LSTM) characteristics can distinguish between deep films and regular movies. On big, balanced, and mixed datasets, our technique is tested by Face-Forensic++ [1], Deepfake Detection Challenge [2], and Celeb-DF [3]. The quality of real-time data is enhanced by this. We prove that our technology consistently outperforms the competitors. Computer vision, Res-Next Convolution neural network, RNN, and LSTM are some of the index phrases.
使用 LSTM 和 resnext-50 进行深度伪造检测 1
随着计算能力的提高,借助深度学习算法,"深度伪造"(即看起来像是真人制作的电影)变得越来越可行。政治不稳定、恐怖主义、报复性色情或勒索都可能成为这些栩栩如生的 "深度伪造 "的动机。这项研究提出了一种新颖的深度学习系统,它可以识别真实电影和人工智能生成的电影之间的区别。我们开发了一种方法,可以自动识别深度假冒替换和重演。我们在用人工智能对抗人工智能。我们系统的 Res-Next 卷积神经网络可检索帧级别的属性。利用长短期记忆(LSTM)特性训练的递归神经网络(RNN)可以区分深度电影和普通电影。在大型、平衡和混合数据集上,我们的技术通过了 Face-Forensic++ [1]、Deepfake Detection Challenge [2] 和 Celeb-DF [3] 的测试。实时数据的质量因此得到了提高。我们证明,我们的技术始终优于竞争对手。计算机视觉、Res-Next 卷积神经网络、RNN 和 LSTM 是其中的一些索引词组。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信