ComputingVision ScienceComputer VisionEdge ComputingCoding TheoryData CompressionDigital Image ProcessingError-correcting CodesImage CompressionImage ProcessingContext-based CompressionEncoder-decoder ArchitecturesImage EncodingLossless Image CompressionLossless Video CompressionResidual CodingVideo Compression
We design an edge-adaptive predictor for lossless image coding. The predictor adaptively weights a four-directional predictor together with an adaptive linear predictor based on information from neighbouring pixels. Although conceptually simple, the performance of the resulting coder is comparable to state-of-the-art image coders when a simple context-based coder is used to encode the prediction errors.
5