Doctoral Thesis: Privacy-Preserving Video Analytics
As video cameras have become pervasive in public settings and accurate computer vision has become commonplace, there has been increasing interest in collecting and processing data from these cameras at scale (“video analytics”). While these trends enable many useful applications (such as monitoring the mobility patterns of cars and pedestrians to improve road safety), they also enable detailed surveillance of citizens at an unprecedented level. Unfortunately, existing solutions fail to practically resolve this tension between utility and privacy, as they rely on perfect detection of all private information in each video frame–an elusive requirement.
In this thesis, we present Privid, a privacy-preserving video analytics system that aims to provide both a meaningful guarantee of privacy and an expressive query interface that is amenable to analysts. In particular, Privid’s privacy definition does not require perfect detection of private information, and its query interface allows analysts to provide their own arbitrary (untrusted) ML processing models. As a result, Privid may be a practical way forward for video analytics to progress while assuring that it cannot aid mass surveillance.
- Date: Friday, May 12
- Time: 3:00 pm - 4:30 pm
- Category: Thesis Defense
- Location: 32-G449
Additional Location Details:
Thesis Supervisor: Hari Balakrishnan
Zoom link is available; contact the doctoral candidate for the link at firstname.lastname@example.org