sTrack: Secure Tracking in Community Surveillance

We develop a system to track objects across multiple cameras without sharing any visual information be-tween two cameras except whether an object was seen by both. To achieve this challenging privacy goal, we leverage recent advances in secure two-party computa-tion and multi-camera tracking. Starting from two dis-tance-metric learning techniques that are foundational for many computer vision tasks, we derive a new tech-nique that is more suited for secure computation be-cause it increases the computation that cameras do locally and simplifies what is done jointly. At the same time, the tracking accuracy of our technique is similar or better than the original techniques. We implement it using a new Boolean circuit for secure tracking. Experiments using real datasets show that the performance overhead of secure tracking is low, adding only a few seconds over non-private tracking.