Image Provenance Analysis

Abstract

The literature of multimedia forensics is mainly dedicated to the analysis of single assets (such as sole image or video files), aiming at individually assessing their authenticity. Different from this, image provenance analysis is devoted to the joint examination of multiple assets, intending to ascertain their history of edits, by evaluating pairwise relationships. Each relationship, thus, expresses the probability of one asset giving rise to the other, through either global or local operations, such as data compression, resizing, color-space modifications, content blurring, and content splicing. The principled combination of these relationships unveils the provenance of the assets, also constituting an important forensic tool for authenticity verification. This chapter introduces the problem of provenance analysis, discussing its importance and delving into the state-of-the-art techniques to solve it.

Unveiling Scientific Articles from Paper Mills with Provenance Analysis

“…As the outcome of researching and developing a computer-aided solution to automatically detect scientific manuscripts from paper mills, we propose a new forensic method extending provenance analysis 7 articles. This is needed because these documents depict imagery that does not follow the usual distribution and nature of figures previously treated by provenance analysis (e.g., microscopic material and Western blots, instead of natural scenes and everyday objects).…”

IdProv: Identity-Based Provenance for Synthetic Image Generation (Student Abstract)

“…To detect identity leakage using existing matching approaches, we require some detectable identity component of the real face signal in the synthetic face. To investigate these concerns further, we examine the provenance (Moreira et al 2022) of this leakage for synthetic face images with composite images. Generally, image provenance for complex manipulated images i.e., query, aims to analyze the evolution of the query image content from its source images (ones that donated content).…”