A+ A- Accessibility
Search

DéjàVu: Feature-Space-Time Coherence from Heterogeneous Data for Media Integrity Analytics and Interpretation of Events

Overview

In this research project, we focus on synchronizing specific events across space and time (X coherence), fact-checking, and mining people, objects, and content of interest from heterogeneous sources, including – but not limited to – the internet, social networks and media. and surveillance footage. For this, we seek to take advantage of information from various media sources and synchronize the multiple textual and visual information around the position of an event or object, as well as order them in order to allow a better understanding of what happened before, during and after after the event. After automatically organizing the data and understanding the order of facts, we can design and apply solutions to mine people or objects of interest for analysis and tracking of suspects, fact-checking or even understanding aspects of the nature of the analyzed event. Furthermore, by exploring the possible connections that exist between different pieces of information, we intend to design and develop media integrity analysis tools to suggest the occurrence of counterfeits, the presence of sensitive content (e.g. violence and child pornography) and patterns of dissemination of objects of online multimedia . With the increasing complexity of digital crimes and terrorist threats, in addition to the spread of false news, our main objective is to use the solutions developed to help answer four important questions in forensic expertise: “who”, “under what circumstances”, “why”. what” and “how”, identifying characteristics and circumstances in which an event occurred.

Duration

2017 – 2022.

Support

Research Support Foundation of the State of São Paulo (FAPESP).

coordinator

Prof. Anderson Rock.