An extension of shape-adaptive DCT (SA-DCT) towards DC separation and Delta DC correction

In the framework of the ongoing MPEG-4 standardization phase of ISO/IEC Sikora and Makai suggested a shape-adaptive DCT algorithm (SA-DCT) for efficient and low-complexity coding of arbitrarily shaped video objects. Referring to this algorithmic approach, the objective of the underlying contribution is to emphasize a performance limitation of conventional SA-DCT definition and to propose an extended version ( Delta DC-SA-DCT) which avoids this restriction in general. This modification considerably improves coding efficiency and can easily be implemented in existing SA-DCT tools.