DCT information recovery of erroneous image blocks by a neural predictor
Francesco G. B. De Natale, Cristian Perra, Gianni L. Vernazza · IEEE Journal on Selected Areas in Communications · 2000
In this paper, the problem of DCT information recovery in the transmission of coded visual data over packet networks is addressed. The loss of a packet conveying coded block data leads to the unsuccessful reconstruction of the relevant area, with consequent degradation of the received image quality. The proposed method allows recovery of a subset of the missing DCT coefficients sufficient to achieve good reconstruction quality of the lost block, based on the available surrounding information. To this purpose, a neural predictor was carefully designed and suitably trained with an appropriate set of synthetic and natural patterns. An extensive testing phase, performed on a large set of images with different frequency characteristics, revealed that the method provides very good reconstruction capabilities.