Evaluating the performance of Fake Face Detection in Forensics

Abstract

Fake faces can be used to disseminate misinformation in an era where digital manipulation iscommonplace. A major
concern in many fields, including forensics, is the proliferation of fake images, particularly those with phoney faces, due to the quick
development of image manipulation technology. The development of techniques for detecting fake faces has attracted a lot of attention
as a responseto this challenge. Because AI-generated fake faces can be used maliciously, the widespread use of these images—of which
thereare over 98%—raises security concerns.
The goal of the project is to use the local binary patterns (LBP)[1] method to determine whether an image is real or fake.It then
assesses the effectiveness of fake face detectiontechniques in the context of forensic applications to identify manipulated or synthesised
facial images.