An image decoding method according to the present document comprises the steps of: acquiring image information through a bitstream, the image information including entry point-related information; deriving at least one entry point in the current picture on the basis of the entry point-related information; and decoding the current picture on the basis of the at least one entry point, wherein the entry point-related information includes an entry point offset presence flag, offset length information, and entry point offset information, the entry point offset presence flag being related to at least one of the number of tiles in the current picture or the number of tile columns in the current picture. Through this method, the amount of data required to be signaled for video/image coding can be reduced, and coding efficiency can be increased.