Research Of The Error Concealment Technology Based On H.264

The new generation of video com pression coding standard H.264, because of its excellent co mpression perform ance and g reat network adaptive, has been widely applied to video transm ission communication se rvices, but as a result of predictiv e coding and transform coding, the compressed data is very sensitive to channel errors.In orde r to im prove the anti-e rror perf ormance, error concealm ent techno logy becomes the research hotspot in recent years.Firstly, the introduction of the research on error concealment techniques at home and abroad is given, and then we analyze and discuss the history of H.264 standard and its new error r esilience tools. Finally , an alysis th e sp atial and te mporal erro r concealment algorithm used in the official H.264 testing platform JM18.0. We design a new error concealment scheme. The main contents of this paper include:(1)Spatial error conc ealment: according to the analysis of bilin ear interpolation method using in JM18.0, we find that the algo rithm has good ef fect in flat area, but has two obvious shortcomings: first, lost the edge information; second: cannot reflect the g ray m utation. This paper provides an error concealm ent algorithm , which combined with the nearest boundary priority sp atial interpolati on and directional interpolation method. After the simulation test and the results analysis w e found that both the subjective and objective performance are better than the JM18.0 algorithm.(2) Temporal error concealment, first we describes some common Temporal error concealment algorithms, then we focusing on analyzing the BMA algorithm using in JM18.0, the BMA algorithm combines spatial correlation and temporal correlation in video pictures and widely used in tem poral error concealment, but als o has som e shortcomings: for example, when the lost block has boundary or the MV has great changes, this method will be invalidated. This paper presents a new algorithm which combines the improved boundary matching algorithm and the bigger reference set, the test results show that this algorithm both in the objective quality and visual effect are better than the BMA algorithm.