Proposal / Submission Type
Peer Reviewed Paper
Abstract
With a growing number of images uploaded daily to social media sites, it is essential to understand if an image can be used to trace its origin. Forensic investigations are focusing on analyzing images that are uploaded to social media sites resulting in an emphasis on building and validating tools. There has been a strong focus on understanding active manipulation or tampering techniques and building tools for analysis. However, research on manipulation is often studied in a vacuum, involving only one technique at a time. Additionally, less focus has been placed on passive manipulation, which can occur by simply uploading an image to a social media site. This research plots the path of an image through multiple social media sites and identifies unique markers in the metadata that can be used to track the image. Both Facebook and Twitter were utilized on both phone and web applications to fully understand any differences between direct and secondary uploads. A full metadata analysis was conducted including histogram and size comparisons. This paper presents several differences and unique metadata findings that allow image provenance to be traced to an original image. This includes a review of IPTC, ICC, and EXIF metadata, ICC profile and Color Profile Description, Encoding Processes, Estimated Quality Values as well as compression ratios. A checklist of variables is given to guide future evaluations of image provenance.
Scholarly Commons Citation
Lint, Gianna S. and Karabiyik, Umit, "Detection of Overlapping Passive Manipulation Techniques in Image Forensics" (2022). Annual ADFSL Conference on Digital Forensics, Security and Law. 4.
https://commons.erau.edu/adfsl/2022/presentations/4
Included in
Aviation Safety and Security Commons, Computer Law Commons, Defense and Security Studies Commons, Forensic Science and Technology Commons, Information Security Commons, National Security Law Commons, OS and Networks Commons, Other Computer Sciences Commons, Social Control, Law, Crime, and Deviance Commons
Detection of Overlapping Passive Manipulation Techniques in Image Forensics
With a growing number of images uploaded daily to social media sites, it is essential to understand if an image can be used to trace its origin. Forensic investigations are focusing on analyzing images that are uploaded to social media sites resulting in an emphasis on building and validating tools. There has been a strong focus on understanding active manipulation or tampering techniques and building tools for analysis. However, research on manipulation is often studied in a vacuum, involving only one technique at a time. Additionally, less focus has been placed on passive manipulation, which can occur by simply uploading an image to a social media site. This research plots the path of an image through multiple social media sites and identifies unique markers in the metadata that can be used to track the image. Both Facebook and Twitter were utilized on both phone and web applications to fully understand any differences between direct and secondary uploads. A full metadata analysis was conducted including histogram and size comparisons. This paper presents several differences and unique metadata findings that allow image provenance to be traced to an original image. This includes a review of IPTC, ICC, and EXIF metadata, ICC profile and Color Profile Description, Encoding Processes, Estimated Quality Values as well as compression ratios. A checklist of variables is given to guide future evaluations of image provenance.
