China
Africa
Explore chapters and articles related to this topic
DeepFake Face Video Detection Using Hybrid Deep Residual Networks and LSTM Architecture
Published in Gaurav Jaswal, Vivek Kanhangad, Raghavendra Ramachandra, AI and Deep Learning in Biometric Security, 2021
Semih Yavuzkiliç, Zahid Akhtar, Abdulkadir Sengür, Kamran Siddique
DeepFakes detection could be considered like a two-class classification problem in which salient characteristics from the given face sample is extracted to be then fed to a classification scheme to attain the binary outcome: DeepFakes or Benign. DeepFakes detection techniques can be broadly categorised into three main classes: textural-, inherent attribute degradations-, and DL-based methods.
Deep Neural Networks
Published in Adrian A. Hopgood, Intelligent Systems for Engineers and Scientists, 2021
Autoencoders are a type of generative network that can be seen as “creative”, in the sense of producing something new at its output (Hinton and Salakhutdinov 2006). They have attracted particular interest for creating so-called deepfake images, which are convincing fake images that have been derived from genuine ones using deep learning (Güera and Delp 2018).
Introduction to DeepFake Technologies
Published in Loveleen Gaur, DeepFakes, 2023
Loveleen Gaur, Saurav Mallik, Noor Zaman Jhanjhi
DeepFake is a collection of “deep learning” and “forgery,” which employs DL algorithms to modify images, acoustic, and video to generate a synthetic/phony media. It is a non-autonomous process that applies AI algorithms to subject matter, producing doctored images, video, and audio.
Harnessing AI for business development: a review of drivers and challenges in Africa
Published in Production Planning & Control, 2022
Joseph Amankwah-Amoah, Yingfa Lu
In recent times, it has become increasingly apparent that technological bias or biases are often built into AI systems. For instance, facial-recognition systems that work less well or are inaccurate for darker-skinned people, and more so darker-skinned women, are likely to lead to false conclusions when used in business processes and decisions (NPR 2019). Indeed, in many sectors, voice-recognition systems tend to perform poorly with people with accents. In addition, some systems also perform better in detecting male-sounding voices relative to female-sounding voices (NPR 2019). It has also been demonstrated that some emotion-detection tools also tend to assign more negative emotions to black men’s faces than white men’s faces (NPR 2019). For small and weak nations, the ‘commercially available AI applications ranging from ‘deepfakes’ to lethal drones’ also possess threats not only to nation states but also provide opportunities for criminals and illegitimate activities to manifest (Schmidt et al. 2021). Thus, deepfakes pose a risk to accountability and reliability of information, and spread of misinformation to the public.
Recent advances in artificial intelligence for video production system
Published in Enterprise Information Systems, 2023
YuFeng Huang, ShiJuan Lv, Kuo-Kun Tseng, Pin-Jen Tseng, Xin Xie, Regina Fang-Ying Lin
Deepfakes are a form of image recognition technology that has made significant contributions to video production and editing. Deepfakes, short for ‘deep learning’ and ‘fake’, utilise artificial intelligence (AI) and deep neural networks (DNNs) to generate hyper-realistic but fabricated content. By training DNNs through deep learning techniques, deepfakes can automatically merge, replace, and superimpose images, audio, and video onto targeted videos, eliminating the need for manual editing or post-production modifications Kietzmann et al. (2019). This sets deepfakes apart from existing methods such as computer-generated imagery (CGI) or Adobe’s Photoshop.
Digital earth: yesterday, today, and tomorrow
Published in International Journal of Digital Earth, 2023
Alessandro Annoni, Stefano Nativi, Arzu Çöltekin, Cheryl Desha, Eugene Eremchenko, Caroline M. Gevaert, Gregory Giuliani, Min Chen, Luis Perez-Mora, Joseph Strobl, Stephanie Tumampos
What is new in our society is the phenomenon of ‘fake data’ (i.e. fake news, fake pictures, fake videos, etc.) fuelled by the latest developments in AI and ML. Now that the technology offers the possibility to create digital fakes quite easily, next-generation AI is threatening to take internet fakery to a dangerous new level. ‘Deepfake’ technology uses sophisticated AI to create video and audio that impersonates real people (Westerlund 2019). The technology is already in use, and if left unchecked, it could lead us to start doubting everything we watch and hear online.