This awardee has received supplemental funding. This award detail page includes information about both the original award and supplemental awards.
Description of original award (Fiscal Year 2009, $150,000)
This project will develop software to assist computer forensic specialists in cases that involve searching for images of humans, such as victims of child pornography. The current state-of-the-practice requires that a computer forensic specialist manually sift through potentially hundreds of thousands of images that are often found on a seized harddrive. The software that the University of Rhode Island produced is very effective at automatically detecting likely pornography and marking those image files for subsequent investigation.
This project will:
' Refine the current detection of general pornography to detect only likely child pornography;
' Integrate the child pornography detection into a Windows forensic boot disk for triage;
' Integrate the child pornography detection into FTK, EnCase, and X-Ways law enforcement analysis tools in a uniform way.
By automating the identification of child pornography in processes and tools familiar to law enforcement, the resulting software will drastically reduce the time required to search for child pornography and generally improve the process for investigators. CA/NCF
This project will to continue development of a software tool to assist computer forensic analysts in searching for images of humans, such as victims of child pornography. The current state-of-the- art requires that a computer forensic analyst sift through the potentially hundreds of thousands of images found on a computer. The proposed technology will automate that process. It is also intended to provide the capability to detect child pornography automatically in video, which is an increasingly prominent form of distribution.
This project has produced RedLight, a software tool that can quickly scan a computer for porn. It has been released free to law enforcement. It has been shown to be effective at automatically detecting likely pornography in both images and video files and marking those files for subsequent investigation. The proposed effort will focus on two objectives. The first is continued development of child detection software with a refined technique to addresses the imprecision the applicant has found when trying to extract child features from images. The second is to repackage RedLight so that it can act as a triage tool during investigation as well as a forensic tool.