Media Revealr: A Social Multimedia Monitoring And Intelligence System For Web Multimedia Verification

Modern online social networks, such as Twitter and Instagram, are nowadays important sources for publishing information and content around breaking news stories and incidents related to public safety, ranging from natural disasters and aeroplane accidents to terrorist attacks and industrial accidents. A crucial issue regarding such information and content is the extent that they can be relied upon and used for improving the situational awareness and operational capabilities of decision makers. Given the proliferation of noisy, irrelevant and fake content posted to such platforms, two important requirements for systems supporting the information access needs in incidents, such as the ones described above, include the support for understanding the "big picture" around the incident and the verification of particular pieces of posted content. To this end, we propose Media REVEALr, a scalable and efficient content-based media crawling and indexing framework featuring a novel and resilient near-duplicate detection approach and intelligent content- and context-based aggregation capabilities (e.g. clustering, named entity extraction). We evaluate the system using both reference benchmark datasets as well as datasets collected around real-world incidents, and we describe the ways it contributes to the improvement of the situational awareness and journalistic verification in breaking news situations, like natural disasters.