Frame Loss Concealment for Multiview Video Transmission Over Wireless Multimedia Sensor Networks

Abstract

In multiview video (MVV) transmission over wireless multimedia sensor networks (WMSNs), a highly compressed video stream is very susceptible to whole-frame losses due to wireless channel errors. Currently, most frame concealment methods can only handle 2-D video or 3-D video with depth, since a lost motion vector can be recovered utilizing the relationship between the color frames and their corresponding depth map. However, depth generation is infeasible in the scenario of WMSNs due to energy constraints and computational limitations of low-power video sensor nodes. In this paper, we tackle the problem of concealing lost whole frames for multiview 3-D video transmission over WMSNs. We propose a new technique, which is capable of dealing with this important problem without the aid of depth map. Experimental results are presented to demonstrate that the proposed algorithm is able to effectively conceal frames of MVV and significantly outperforms other techniques in both objective and subjective quality measurements.