N (N≥2) time stretch units receive N channels of second image data which are obtained by dividing each line of first image data by N and are composed of a first number of bits. Each N time stretch unit arranges pixels of the N channels in the order of the pixels of the corresponding line of the first image data within a period which is obtained by stretching a one-line period of the second image data. The N time stretch units generate third image data including N lines of the first image data. N encoders encode each coding target pixel of the third image data outputted from the N time stretch units using a difference between the target pixel and a peripheral pixel to generate encoded data composed of a second number of bits which is smaller than the first number of bits.