Testing the Working Efficiency of Open-Source Tools IrfanView and FotoForensics against Deep-Fakes Detections in Images

Abstract

DeepFake and image manipulation pose major challenges about the authenticity of any digital media such as image and video. Thus, demanding more efficient tools for their detection. This research paper provides a comprehensive analysis of two open-source tools: FotoForensics and IrfanView, about their efficiency in detecting DeepFake and image manipulation. Furthermore, this study addresses these tools capabilities and their limitations in addressing challenges. To conduct our research work, different datasets were gathered from Kaggle and Google search. Our research mainly focuses on application of different features of these tools such as metadata analysis, ELA, hidden pixels, negatives, histograms and edge detection. By subjecting the gathered datasets to these features, we determine the working efficiency of these tools. Through our findings, we aim to give some valuable insights for practitioners and researchers towards enhancing DeepFake and image manipulation detection methods.