An image decoder including: a demultiplexer extracting motion vector information, quantized DCT coefficients, and rounding method information; a dequantizer dequantizing the quantized DCT coefficients to obtain DCT coefficients; an inverse DCT converter performing inverse DCT conversion on the DCT coefficients to obtain an error image; a prediction image synthesizer synthesizing a prediction image by performing motion compensation using the motion vector information, the rounding method information, and a reference previously decoded image; and an adder adding the prediction image to the error image to obtain a decoded image. Motion compensation is performed with half-pixel accuracy and bilinear interpolation to calculate intensity values of chrominance or luminance at points where no pixels actually exist in the reference image, using positive and negative rounding methods. Bilinear interpolation is performed using a rounding method specified by the rounding method information included in coded information of the currently decoded image.