This paper proposes and evaluates a promising technique used for liver segmentation that can be used as the first step in liver treatment planning. It is based on texture feature analysis obtained from 3D DCT block coefficients, support vecto