Picture coding using weighted predictions in the transform domain

摘要

An image encoder utilises a transformation operating between a spatial domain and transform domain, for example the Discrete Cosine Transform (DCT) domain. The encoder further employs the steps of forming a prediction; subtracting the prediction to form a difference; and quantising the difference in a transform domain. Furthermore the prediction is formed in the transform domain and is weighted. The weighting may be carried out using a matrix so that different coefficients can be weighted differently. Similarly the weighting can vary between image blocks. The method is expected to be of particular use for video pictures.