METHOD OF COMPRESSING IMAGES Russian patent published in 2016 - IPC G06T9/00 H03M7/30 H04N19/25 

FIELD: physics.

SUBSTANCE: method of compressing a digital colour image comprises: dividing into colour components, each being a digital halftone image represented by g-bit binary numbers; dividing the digital halftone image into bit binary images; selecting a vicinity for each image element; for the most significant bit binary images containing large image elements, applying element prediction procedure 1 based on correlation coefficients on the horizontal and the vertical and the theory of conditional Markovian processes and removing correctly predicted elements; for the least significant bit binary images containing small image details and the background, which is similar in structure to white Gaussian noise, applying procedure 2 which includes selecting the regions containing image details and applying procedure 1 thereto; removing the regions containing the background; encoding the bit stream.

EFFECT: faster operation and efficiency of compressing images, eliminating image integrity loss and maintaining the contrast of boundaries between different image objects.

2 cl, 5 dwg