Video coding using lapped orthogonal transform

An orthogonal transform is an essential technique for reducing the spatial redundancy of image signals. In the orthogonal transforms, Discrete Cosine Transform (DCT) has been most successfully applied to image codings, since it provides a high compensation ratio. However, DCT coding produces discontinuous noise at block boundaries called 'blocking effect' which degrades low bitrate coded images. To reduce the blocking effect, Lapped Orthogonal Transform (LOT) has been proposed. This paper proposes an intrafield/interframe adaptive LOT and an LOT domain decimation to apply LOT to a scalable/compatible video coding. The proposed method is compared with DCT-based method in terms of coding efficiency, picture quality and extendability to the scalable/ compatible coding. The advantages of LOT over DCT are less blocking effect, higher coding efficiency and less aliasing noise in the scaled images obtained by transform domain decimation.