2000 character limit reached
Non-Semantic Evaluation of Image Forensics Tools: Methodology and Database
Published 4 May 2021 in eess.IV and eess.SP | (2105.02700v1)
Abstract: With the aim of evaluating image forensics tools, we propose a methodology to create forgeries traces, leaving intact the semantics of the image. Thus, the only forgery cues left are the specific alterations of one or several aspects of the image formation pipeline. This methodology creates automatically forged images that are challenging to detect for forensic tools and overcomes the problem of creating convincing semantic forgeries. Based on this methodology, we create the Trace database and conduct an evaluation of the main state-of-the-art image forensics tools.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.