Extracting and Encoding Mutually Sensed Region in Wireless Multimedia Sensor Networks to Remove Redundancy

Abstract

A wireless multimedia sensor networks (WMSNs) an emerging field of sensor networks has wide variety of multimedia applications. Due to its ever increasing range of visual applications data redundancy and correlated information distribution is major issue in WMSNs. In this paper we have introduced a technique to remove redundant information. Extracting mutual sensed region is the core of proposed approach. To extract mutual sensed region first scale invariant feature transform (SIFT) is applied for selecting control points in correspondence with two views. After extracting control points affine image registration is applied to find mutual sensed region, and affine matrix is used to assist node for first frame onward communication. The proposed approach is implemented and tested in MATLAB and it shows decrease in size of frames as compared to H.264 encoding standard.