Abstract
A method and an apparatus for cost calculation in decimal motion estimation are provided, which have the advantages of high video compression quality and reduced amount of calculation. The method comprises the following steps. Firstly, perform interpolation on a current block to get an interpolation result at a position corresponding to a decimal motion vector. Secondly, calculate a cost according to data at integer point positions of a reference frame corresponding to the current block and the decimal motion vector, and according to the interpolation result.