Abstract:
Transmission of 3D video content over IP network faces challenges caused by network parameters such as packet loss, delay and jitter. Also, its quality is impacted by limited bandwidth and less storage. 3D video contains depth information which has more impact on the quality than 2D video and this is due to network impairment. End user requires high quality to receive their multimedia services and application, therefore, this research proposed an efficient hybrid concealment scheme that aims to minimize 3D video quality degradation as a result of packet loss on color and depth map of stereoscopic video sequences. The proposed hybrid concealment scheme focusses on spatial prediction algorithm using nearest neighboring macroblock in slice by exploiting the K Nearest Neighbors (KNN) regression method to conceal missing packets in slice. The proposed scheme is based on two resilience techniques in H264/AVC coding. These techniques are slice groups which is also known as Flexible Macroblock Ordering (FMO) and Redundant Slices (RS). These techniques control errors that are spread spatially on slices. The subjective tests and the simulation results have shown that the proposed hybrid concealment scheme achieved better in 3D video quality than the two previous methods in the literature (H264/AVC Frame Copy and Depth Assisted)
