METHOD AND DEVICE FOR COMPRESSING VIDEO INFORMATION FOR TRANSMISSION OVER COMMUNICATION CHANNELS WITH VARYING THROUGHPUT CAPACITY AND STORAGE IN DATA STORAGE SYSTEMS USING MACHINE LEARNING AND NEURAL NETWORKS Russian patent published in 2019 - IPC G06F16/70 H03M7/30 H04N21/238 

FIELD: computer equipment.

SUBSTANCE: invention relates to the computer equipment. A method of compressing video information in digital form for transmission thereof over non-stationary communication channels of a data transmission network or storage in data storage systems, in which a color or black-and-white tone video stream is subjected to preprocessing and conversion into a black-and-white sketch form, is encoded taking into account specificity, packaged and sealed multiplexer with additional service data flow transmitted or placed in storage device, at output is restored by decoder in sketch form after decompression by demultiplexer, depacketizing of compressed video information and service data, wherein sketch-flow frames with low dynamics of objects in the scene are excluded from transmission using an interpolator and threshold selection of frames in the encoder; wherein the main encoded sketch stream is supplemented with service data packets on missed frames and a separate stream of first frame-hints generated when detecting scene change, if the scene change information and the first frame of the new scene are not provided for training the INN engine.

EFFECT: high efficiency of processing a video stream for transmitting data in a network environment.

10 cl, 4 dwg