A method for decoding an image by a decoding apparatus comprises the steps of: obtaining merge with motion vector difference (MMVD) flag information of a current block from a bit stream; determining whether or not the current block is a current picture referencing (CPR) coding block; if the current block is a CPR coding block, deriving base motion information; deriving motion information of the current block on the basis of the base motion information; generating prediction samples of the current block on the basis of the motion information; and generating reconstruction samples of the current block on the basis of the prediction samples, wherein the MMVD flag information indicates whether or not motion information is derived when the MMVD is applied to the current block, and the CPR coding block is a block coded by using a current picture including the current block as a reference picture.