Long-Term Anomaly Detection in Camera Networks
R4-A3

Automatic detection of anomalous tracks and scenes at widely different time scales

Relevance to the Homeland Security Enterprise

  • Thousands of unmanned cameras at DHS locations (airports, transit stations, border crossings)
  • Automatic algorithms needed to flag possible anomalies for further human inspection
  • Low tolerance for false alarms; extremely low tolerance for misses
Project Leader
  • Richard Radke
    Professor
    Rensselaer Polytechnic Institute
    Email

Students Currently Involved in Project
  • Matthew Reome
    RPI
  • Mahdee Jameel
    RPI
  • Srikrishna Karanam
    RPI
  • R4-A3
  • Austin Li
    RPI