An effective post-processing algorithm for error recovery In noise-corrupte
d JPEG bit streams is presented. The good performances provided in terms of
perceptual quality and real-time working capability makes the algorithm hi
ghly suitable for applications such as outdoor video-surveillance, characte
rised by severe constraints on image quality and processing time.