Title: Deepfake detection and localisation based on illumination inconsistency

Authors: Fei Gu; Yunshu Dai; Jianwei Fei; Xianyi Chen

Addresses: Nanjing University of Information Science and Technology, Nanjing, 210044, China ' Nanjing University of Information Science and Technology, Nanjing, 210044, China ' Nanjing University of Information Science and Technology, Nanjing, 210044, China ' Nanjing University of Information Science and Technology, Nanjing, 210044, China

Abstract: The rapid development of image synthesis technology has encouraged the spread of some fake news, making people gradually lose trust in digital media. The compression in the process of image propagation brings a major challenge to the existing face forgery detection method. In this paper, we propose a multi-task Deepfake detection method according to the motivation of illumination inconsistency between tampered and non-tampered areas. Specifically, we trained a Siamese network as a feature extractor to estimate the illumination, then distinguish the face image and predict the forged region through a U-shaped network. Our method has achieved great accuracy in classification tasks and can still maintain good performance in compressing data. In addition, we can also show the intensity of tampering while locating the forged area.

Keywords: Deepfakes; illumination estimation; Siamese network; UNet; image manipulation detection; image forensics; face spoof detection; convolution neural network; artificial intelligence security; Deepfake detection; face forensics; deep learning.

DOI: 10.1504/IJAACS.2024.139383

International Journal of Autonomous and Adaptive Communications Systems, 2024 Vol.17 No.4, pp.352 - 368

Received: 13 May 2022
Accepted: 04 Jul 2022
Published online: 02 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article