An image compression unit includes a block divider 11 that divides image data into a plurality of blocks, a DCT portion 12 that subjects the image data of each block to DCT, a quantization factor output portion 15 that outputs a quantization factor as a numerical value indicative of fineness of quantization, a correction value computation portion 16 that computes a correction value that is determined uniquely according to a position of a block of image data, a quantizer 13 that quantizes the data having been subjected to DCT, based on a value obtained by multiplying the correction value by the quantization factor, and an encoder 14 that encodes the quantized data and outputs compressed data. This configuration makes it possible to provide an image compression/decompression unit that is capable of varying a compression ratio in a frame, without encoding compression parameters of each block.