Binocular video compression method and device based on deep learning, and readable medium

Abstract

The invention discloses a binocular video compression method and device based on deep learning and a readable medium, and relates to the field of video processing, and the method comprises the steps: constructing and training a binocular video compression model, and obtaining a trained binocular video compression model, the method comprises the following steps: respectively acquiring an original frame of a current moment and an original frame of a previous moment of a left viewpoint and an original frame of a current moment and an original frame of a previous moment of a right viewpoint in a binocular video to be compressed, and inputting the original frames into a trained binocular video compression model; and reconstructing to obtain a reconstructed frame sequence of the left viewpoint and a reconstructed frame sequence of the right viewpoint in an iteration mode, and obtaining a compressed binocular video. According to the method, the problems of low redundancy utilization rate and low compression performance between viewpoints of the existing binocular video are solved.

Images