Resilient Video Coding using Difference Expansion and Histogram Modification

Abstract:
Recent advances in multimedia technology have paved the way to the development of several applications, including digital TV broadcasting, mobile TV, mobile gaming and telemedicine. Nonetheless, real time multimedia services still provide challenges as reliable delivery of the content cannot be guaranteed. The video compression standards incorporate error resilient mechanisms to mitigate this effect. However, these methods assume a packet-loss scenario, where corrupted slices are dropped and concealed by the decoder.

This paper presents the application of reversible watermarking techniques to facilitate the detection of corrupted macroblocks. A variable checksum is embedded within the coefficient levels and motion vectors, which is then used by the decoder to detect corrupted macroblocks which are concealed. The proposed method employs difference expansion to protect the level values while histogram modification was employed to protect the motion vectors. Unlike previous published work by the same author, this scheme does not need the transmission of side information to aid the recovery of the original level and motion vector values. Simulation results have indicated that significant gains in performance can be achieved over the H.264/AVC standard.