A method for compressing image data having pixels arranged like a matrix in a line direction and a sub scan direction orthogonal to the line direction is provided herein. The disclosed method allows for a reduction in memory capacity necessary for the expansion of the data. The method includes the steps of dividing the image data in the line direction and in the sub scan direction into blocks covering a plurality of lines, determining a representative value about density of pixels included in each of the blocks, determining additional information for each line in each of the blocks or for a plurality of lines whose number is smaller than the total number of lines of the block and creating compressed data by using the representative value and the additional information.