Visual Sensor Network Processing and Preventative Steganalysis

Background

Most well known measures to protect WVSNs, to date, have focused on the problem of providing privacy in vision-rich systems. Lo et al. (Lo, Wang & Yang, 2005) introduce an automated homecare monitoring system for the elderly named UbiSense where image processing is conducted directly at the camera to convert visual data directly into abstractions that reveal no personal information and hence protect the privacy of the monitored individuals. Fidaleo et al. (Fidaleo, Nguyen & Trivedi, 2004) introduce the Networked Sensor Tapestry (NeST) architecture designed for the secure sharing, capture, and distributed processing and archiving of multimedia data. They introduce the notion of “subjective privacy” in which processing of raw sensor data is conducted to remove personally identifiable information; thus the behavior, but not the identity of an individual under surveillance is conveyed. The resulting data, approved for public viewing, is communicated in a network that employs the secure socket layer protocol and client authorization for network-level protection. Wickramasuriya et al. (Wickramasuriya, Datt & Mehrotra, 2006) present a privacy preserving video surveillance system that monitors subjects in an observation region using video cameras along with motion sensors and RFID tags. The motion detectors are used to trigger the video cameras on or off, and the RFIDs of the subjects provide authorization information in order to specify which individuals are entitled to privacy and hence have their visual information masked through image processing. More recently, Kundur et al. (Kundur, 2008)(Kundur, Luh, Okorafor & Zourntos, 2008) present the HoLiSTiC (Heterogeneous Lightweight Sensornet for Trusted Visual Computing) framework for WVSN security that exploits secure protocols in a hierarchical directional link communication network to achieve broadband low power communications. A decentralized visual secret sharing approach is used to preserve privacy.