作者: M.-H. Jo , H.-N. Kim , W.-J. Song
关键词: Data compression 、 Error detection and correction 、 Computer vision 、 Video compression picture types 、 Transmission (telecommunications) 、 Signal compression 、 Image processing 、 Macroblock 、 Computer science 、 Block (data storage) 、 Artificial intelligence
摘要: We present a novel error-concealment method for MPEG-2 video decoders. Imperfect transmission of block-based compressed images may result in the loss blocks, degrading quality received signal. Since an I-picture is used as reference picture forward and backward predictions neighbouring pictures, errors are propagated both spatially temporally to P-/B-pictures, thus severely degrade signal, compared with P-/B-pictures. To minimise degradation error concealment has been performed I-pictures through adaptive spatial temporal interpolations respective domains. Using edge information extracted from neighbourhoods erroneous macroblocks, we categorise each macroblock (MB) based on characteristics MBs. The categorised MB reconstructed by or methods selected category-dependent criteria reflecting properties block content. proposed technique solves blocky artefacts blurring problems existing effectively conceals lost blocks Through computer simulations damaged images, show that robust image obtains better subjective than other methods.