Privacy-Preserving Surveillance as an Edge Service Based on Lightweight Video Protection Schemes Using Face De-Identification and Window Masking

Abstract

With a myriad of edge cameras deployed in urban and suburban areas, many people are seriously concerned about the constant invasion of their privacy. There is a mounting pressure from the public to make the cameras privacy-conscious. This paper proposes a Privacy-preserving Surveillance as an Edge service (PriSE) method with a hybrid architecture comprising a lightweight foreground object scanner and a video protection scheme that operates on edge cameras and fog/cloud-based models to detect privacy attributes like windows, faces, and perpetrators. The Reversible Chaotic Masking (ReCAM) scheme is designed to ensure an end-to-end privacy while the simplified foreground-object detector helps reduce resource consumption by discarding frames containing only background-objects. A robust window-object detector was developed to prevent peeping via windows; whereas human faces are detected by using a multi-tasked cascaded convolutional neural network (MTCNN) to ensure de-identification. The extensive experimental studies and comparative analysis show that the PriSE scheme (i) can efficiently detect foreground objects, and scramble those frames that contain foreground objects at the edge cameras, and (ii) detect and denature window and face objects, and identify perpetrators at a fog/cloud server to prevent unauthorized viewing via windows, to ensure anonymity of individuals, and to deter criminal activities, respectively.

Dutkat: A Multimedia System for Catching Illegal Catchers in a Privacy-Preserving Manner

“…Fitwi et al [5] describe a system for masking private information in video frames from surveillance cameras by doing detection and filtering on the edge. In their paper, they discuss the advantages and disadvantages of a cloud-based surveillance system vs. an edge-based approach.…”

Estimating Interpersonal Distance and Crowd Density with a Single-Edge Camera

“…A survey [18] shows the attempts made to make use of some of the aforementioned technologies; they lack clear methods for precisely determining the distance between people in a crowd, though. Besides, previous works that focus on selective surveillance [6,9] and crowd surveillance using drones [13,[19][20][21] could be further developed to be employed for social distance determination, monitoring, and alerting. The machine learning technology has a wide range of applications [22,23]; as a result, it can be adopted to design and build models useful for crowd control, like monitoring social distancing.…”

“…It is one of the important means of meeting the challenges posed by the rising crime rate. Today, with the main goals of ensuring physical security and public safety, there are more than a billion CCTV cameras in use around the globe enabling the law enforcers and security personnel to collect huge amount of information about individuals and follow their activities live [4][5][6][7][8]. It helps to identify law breaking individuals and deters crimes.…”

File System Support for Privacy-Preserving Analysis and Forensics in Low-Bandwidth Edge Environments

“…Similar video analysis must be performed for privacy, e.g., avoiding to show faces or objects that should for some reason be protected. For example, Fitwi et al [52] describe a system for masking private information in video frames from surveillance cameras by doing detection and filtering on the edge. Moreover, D'souza et al [53] describe a similar system that uses object detection for surveillance camera video streams, and whitelists classes of objects that should not be censored.…”