Computer vision models trained on public datasets often exhibit performance drift when deployed in real-world surveillance, diverging from their reported test-set results. This workshop invites papers presenting experimental results from any real-world surveillance or asset-protection application (e.g.,guarding buildings and critical infrastructure), the challenges encountered, and strategies for mitigation on topics including, but not limited to:

The event aims to explore a range of themes, including the following: computer vision, real world, surveillance and applications.