Advanced learning techniques for detection of contraband video

The ease of access, relative anonymity, and the border less nature of the Internet has made child pornography a problem for governments and law enforcement officials around the world. Despite the existence of a great many laws worldwide that prohibit both possession and distribution of such content, it has proven to be an extremely difficult problem to curtail on-line storage and distribution of child pornography.
This project aims to design and implement novel algorithms for detecting contraband video. The project involves an academic researcher as well as an experienced practitioner and an industry specialist. The proposed research is directed at two closely related goals: ({\em i}) identifying whether a given video contains nudity and ({\em ii}) determining whether the video has children as subjects. To achieve these goals, tools and algorithms will be developed based on the PI's prior experience with image data and recent machine learning techniques and 3D reconstruction techniques that can be used with video data.
The algorithms developed to solve these problems will be implemented in a software toolkit and evaluation will be done with realistic test cases developed in collaboration with the Department of Homeland Security, Child Exploitation Investigations Unit (ICE) and Griffeye software. Griffeye is a provider of the world's premier intelligence and visual big data platform for collecting, processing, analyzing, visualizing and managing images and videos. Work done will be made part of the Project VIC initiative providing an API for any tool serving the Project VIC community through the Project VIC Github. Project VIC is a collaborative initiative between law enforcement, industry and non-governmental organizations providing technology, tools, work-flows and a framework to increase law enforcement efficiency in dealing with child exploitation seized data. Griffeye donates their Analyze DI product to law enforcement utilizing the tool for child exploitation investigations and has donated Analyze DI to Project VIC. ca/ncf