This paper investigates the performance of different temporal concealment techniques when incorporated within a long-term motion compensated video codec. In particular, the paper finds a combination of techniques that best recoversthe spatial-temporalcomponentsof a damaged longterm motion vector. It is demonstrated also that accurate recovery of the spatial componentsof the damaged motion vectoris, in general,more importantand morecomplexthan the recovery of the temporal component. Such findingsare explained in view of the properties of the long-term block motion field. The paper, also, proposes a novel long-term multihypothesistemporalconcealmenttechnique where the damaged block is concealed using the averageof a number of candidates from the severalreference frames availableat the decoder. The superior performance of this technique is demonstratedover a range of block error rates.
Mohammed E. Al-Mualla, Cedric Nishan Canagarajah,