Scalable human-machine collaborative image coding method and system, decoder training method

Abstract

The invention discloses an expandable man-machine cooperative image coding method and a coding system. The method is as follows: extract the edge map of each sample picture and vectorize it as a compact representation for driving machine vision tasks; extract key points from the vectorized edge map as auxiliary information; perform entropy encoding on the compact representation and auxiliary information respectively Lossless compression to obtain two code streams; preliminarily decode the two code streams to obtain edge maps and auxiliary information; input the decoded edge maps and auxiliary information into the neural network for forward calculation of the network; Calculate the loss function of the calculation result and the corresponding original image, and backpropagate the calculated loss to the neural network to update the network weight until the neural network converges to obtain a two-way code stream decoder; obtain the edge map and auxiliary information of the image to be processed After coding and compressing, two code streams are obtained; the two-way code stream decoder decodes the received code streams and reconstructs images.

Description

technical field [0001] The invention belongs to the field of image coding, and relates to a scalable human-machine collaborative image coding method and coding system. The invention can simultaneously improve the quality of images under human vision and machine vision. Background technique [0002] In the process of using and spreading digital images, lossy image compression is an indispensable key technology. The traditional lossy image compression scheme transforms the image to obtain a compact representation, and then continues to quantify and entropy encoding for compression, which greatly reduces the overhead of digital images in the process of storage and transmission, and enables digital images to be widely used in daily life. [0003] With the development of computer vision technology, more and more application scenarios need to consider the quality of images under machine vision, that is, images after lossy compression can still maintain comparable performance to lo...

AI Technical Summary

Problems solved by technology

However, traditional lossy image compression schemes are only optimized for human vision and cannot guarantee the quality of machine vision

However, if we only consider compressing the features of machine vision tasks and do not guarantee the recovery and reconstruction of images, we cannot observe them under human vision.

Images