The method and apparatus particularly adapted for reducing the storage capacity of line memories needed for temporarily storing image data for lines until image data for the last line is obtained in digital filtering. Inputted image data are divided into a plurality of blocks, each of which consists of image data obtained from n pixels. The integer n may assume a value of 8, 16, 32, etc. The constituents of one block after another are compressed such that image data consisting of 8 bits per pixel are converted into encoded data consisting of 2 bits per pixel. Data on a dynamic range and an average value for each block are added as a headder to the encoded data. Compressed image data are taken out of a buffer memory after being temporarily stored therein and, together with or separately from compressed image data newly outputted from a compression unit, subjected to image processing such as MTF correction in an arithmetic unit. Image processing to be carried out in the arithmetic unit normally includes expansion and correction to be carried out independently of or simultaneously with each other.