02-09-2017, 03:15 PM
In video communication, masking of distortions caused by transmission errors is important to allow pleasant visual quality and to reduce the propagation of errors. In this article, Denoised Temporal Extrapolation Refinement is introduced as a novel spatiotemporal error concealment algorithm. The algorithm works in two steps. First, concealment of the temporal error is used to obtain an initial estimate. Subsequently, a spatial degradation algorithm is used to reduce the imperfection of temporal extrapolation. For this purpose, the non-local media destructor is used, which is expanded by a spiral scanning processing order and is improved by an adaptation step to take into account the preliminary temporal extrapolation. In doing so, there is a concealment of the spatio-temporal error. By making use of refinement, a visually perceptible average gain of 1 dB over concealment of pure temporal error is possible. With this, the algorithm is also able to clearly overcome other hidden algorithms of spatio-temporal errors.
Recent video encoding standards achieve efficient compression using motion estimation and compensation techniques, discrete cosine transformation, and variable length encoding. H.264 or Advanced Video Coding is the international video coding standard provided jointly by the ITU-T Video Coding Expert Group and the ISO / IEC Moving Pictures Expert Group. Compared with previous standards, H.264 / AVC adopted new coding tools, such as Intra prediction, loop filtering, structured motion estimation tree and compensation and so on. This allows the encoded bit stream to have more information, such as parameter sets, Intra prediction modes and flexible cut-off information. When compressed video bitstream is transmitted through error prone channels such as a packet loss network, it may be corrupted by channel noise. All the information contained in the lost packets will not be available in the decoder and therefore the video reconstruction is of poor quality. Many error resiliency techniques such as re-synchronization marker insertion and forward error correction (FEC) have been proposed for adding lateral information in the coded side stream of the encoder to make it more resistant to channel errors. H.264 / AVC also employs several new error resiliency tools to combat channel errors. However, they still can not guarantee to compensate for channel errors with good quality of rebuild. To reduce quality degradations and improve video quality, error concealment techniques are used on the decoder side as a post-processing module. Error concealment has the advantages of not consuming additional bandwidth as FEC and not introducing retransmission delays as an automatic retransmission request (ARQ). An error detection mechanism is required before hiding the error for the location of the error region in the decoded video.
Recent video encoding standards achieve efficient compression using motion estimation and compensation techniques, discrete cosine transformation, and variable length encoding. H.264 or Advanced Video Coding is the international video coding standard provided jointly by the ITU-T Video Coding Expert Group and the ISO / IEC Moving Pictures Expert Group. Compared with previous standards, H.264 / AVC adopted new coding tools, such as Intra prediction, loop filtering, structured motion estimation tree and compensation and so on. This allows the encoded bit stream to have more information, such as parameter sets, Intra prediction modes and flexible cut-off information. When compressed video bitstream is transmitted through error prone channels such as a packet loss network, it may be corrupted by channel noise. All the information contained in the lost packets will not be available in the decoder and therefore the video reconstruction is of poor quality. Many error resiliency techniques such as re-synchronization marker insertion and forward error correction (FEC) have been proposed for adding lateral information in the coded side stream of the encoder to make it more resistant to channel errors. H.264 / AVC also employs several new error resiliency tools to combat channel errors. However, they still can not guarantee to compensate for channel errors with good quality of rebuild. To reduce quality degradations and improve video quality, error concealment techniques are used on the decoder side as a post-processing module. Error concealment has the advantages of not consuming additional bandwidth as FEC and not introducing retransmission delays as an automatic retransmission request (ARQ). An error detection mechanism is required before hiding the error for the location of the error region in the decoded video.