The number of video surveillance deployments in publicly accessible spaces is continuously increasing. At the same time, the call for technical assistance of human operators in terms of so-called smart video surveillance system has been growing ever louder and is addressed by research and development. With such systems, video analysis algorithms for activity recognition, tracking, and biometric...

Increasing capabilities of intelligent video surveillance systems impose new threats to privacy while, at the same time, offering opportunities for reducing the privacy invasiveness of surveillance measures as well as their selectivity. We show that aggregating more data about observed people does not necessarily lead to less privacy, but can increase the selectivity of surveillance measures. I...

In this paper, we describe EPFL privacy protection tool submitted to the MediaEval 2013 Visual Privacy task. The goal of the task is to obscure faces and other personal items of people in the provided surveillance clips to preserve their personal privacy. In the privacy protection tool, we used a combination of reversible privacy protection filter based on geometric warping transformation, rand...

O VER the past decade, video surveillance has rapidly become a widespread tool for public security. However, with the increasing volume of surveillance data, traditional manual monitoring suffers from human errors and potentially considerable delays in decision making. Moreover, there is an increasing concern of invasion of personal privacy as people's whereabouts and activities are recorded an...

Aiming at the need of personal privacy protection in video surveillance, especially in the traffic surveillance, we discuss in this paper, approaches for preserving privacy for traffic surveillance. We propose three approaches so that the observer can only view a transformed form of surveillance to protect individual privacy. In the first approach, we utilize randomization based techniques usin...