Accession Number : AD1060877


Title :   Machine Learning for Adaptable Heterogeneous Indexing and Search


Descriptive Note : Technical Report,01 Sep 2014,31 Mar 2018


Corporate Author : Carnegie Mellon University Pittsburgh United States


Personal Author(s) : Dubrawski,Artur


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/1060877.pdf


Report Date : 01 Sep 2018


Pagination or Media Count : 20


Abstract : This work has developed new technology that matches content between imagery relevant to a certain domain. The focus of this effort was on a law enforcement application concentrating on Human Trafficking (HT) in order to identify the room and background images that would appear in illicit advertisements and connect them to those available on hotel websites. These images would then be gathered according to hotel chain, and then further refined to individual location which would lead to estimates of GPS coordinates. Contractor also employed their unique facial recognition methods to match individuals across ads. The technology is being validated through transition to multiple law enforcement agencies.


Descriptors :   Facial Recognition , Machine Learning , IMAGE PROCESSING , artificial neural networks , law enforcement , pattern recognition , identification , algorithms , network science , data mining , human trafficking , CRIME


Subject Categories : Information Science


Distribution Statement : APPROVED FOR PUBLIC RELEASE