Lopata Hall, Room 101
"Using Deep Learning Techniques in Image Reconstruction"
Adviser: Mark Anastasio
Due to the nature of this work, an abstract cannot be made public at this time.
"TraffickCam: Crowdsourced and Computer Vision-based Approaches to Fighting Sex Trafficking"
Adviser: Sanmay Das
According to a 2016 research study interviewing victims of sex trafficking, over sixty percent of child sex trafficking survivors were at one point advertised online. These advertisements include photos of the victim often posed provocatively in a hotel room. It is imperative that law enforcement be able to quickly identify where these photos were taken in order to determine the specific locations and the geographic extent where a trafficker moves their victims. In order to determine the hotel in the photos of victims, law enforcement currently perform time consuming manual investigations; for example, they ask individuals who are regular travelers if they recognize the location photographed, and compare the photos to those on travel websites, which can often be out of date or of only the nicest rooms at a hotel with professional photography. We have implemented a system to crowdsource the collection of representative images of hotels around the country using a smartphone application, present different approaches to identifying a hotel room in a query image, and discuss the challenges of implementing such a system for law enforcement nationwide.