A method (200) and apparatus (100) for compressing an image is disclosed. The image is provided as one or more runs of pixels on a scanline basis. The runs of pixels represent one or more graphical objects forming the image. For each scanline of the image, pixel difference values between each adjacent pixels of at least one run of pixels of the scanline is determined. The pixel difference values for the run of pixels are then compared to predetermined difference values. One or more coded bytes are determined for the run of pixels depending on the comparison, each of the coded bytes representing a difference value between at least two adjacent pixels of the run of pixels. The one or more coded bytes are then stored as a compressed representation of the run of pixels corresponding to the scanline of the image.