RedLight an efficient illicit image detection application for law enforcement

Sean P Alvarez, University of Rhode Island


Law enforcement investigators manually sort through hundreds of thousands of images when investigating a hard drive for the presence of illicit images. This is error prone, as many items may be missed due to the volume of images investigated. This is also very time-consuming for the investigator, creating backlogs of cases needing to be processed that can stretch for months or even longer. This thesis presents a user friendly application that is immediately available for use by law enforcement that can be used to increase the efficiency of searching suspect storage media for illicit content in both images and video. By using this application, an investigator can process new investigations in a much shorter period of time than can be done manually, allowing for extra time to get caught up on any backlog of cases. The application is approximately 85% accurate at detecting illicit images, and is able to process at least 38 images per second, which is significantly faster than can be done manually. The application is multi-threaded in such a way that processing speed will go up depending on how many processing cores are available in the machine. The application requires less than 30MB of memory to start, allowing it to run on any relatively modern machine. This enables any department or division of law enforcement to use the application without incurring any costs of upgrading or replacing workstations.^

Subject Area

Sociology, Criminology and Penology|Computer Science

Recommended Citation

Sean P Alvarez, "RedLight an efficient illicit image detection application for law enforcement" (2012). Dissertations and Master's Theses (Campus Access). Paper AAI1508309.