Variable code rate video compression method, system and device and storage medium

Abstract

The invention provides a variable code rate video compression method, system and device and a storage medium, and the method comprises the steps: enabling a reference frame image to pass through a convolutional neural network to obtain a first feature map of a reference frame, obtaining a second feature map of the reference frame through scale zooming and derivable quantization, and carrying out the inverse zooming to obtain a reconstructed reference frame. Processing the prediction frame image through a convolutional neural network to obtain a first feature map of the prediction frame, and performing scale scaling and quantization to obtain a second feature map of the prediction frame; subtracting the second feature map of the prediction frame from the second feature map of the reference frame to obtain a feature map residual error; using a second feature map of the reference frame as prior information, performing lossless entropy coding on a feature map residual error, and transmitting the feature map residual error to a decoding end; and after the feature map residual error and the second feature map of the reference frame are added and inversely scaled, a reconstructed prediction frame is obtained through a convolutional neural network. According to the method, the scale factors are used for scaling the feature maps, so that the feature maps corresponding to different code rate qualities are obtained.

Description

technical field [0001] The invention belongs to the field of image processing, and relates to an end-to-end optimized variable bit rate video compression scheme using time domain and space domain context information, in particular to a variable bit rate video compression method, system, device and storage medium. Background technique [0002] End-to-end optimization is a technique commonly used in image compression and video compression. It uses convolutional neural networks to model compression as a rate-distortion bidirectional optimization problem, which can jointly optimize various modules. The variable bit rate scheme can output output images or video sequences with different quality and different bit rates for the input image or video sequence through a single model. After searching the literature of the prior art, it was found that "DVC: An End-to-End Deep Video Compression Framework" published by Lu et al. at the "IEEE Conference on Computer Vision and Pattern Recogn...

AI Technical Summary

Problems solved by technology

The input of the optical flow network at the encoding end is the reconstructed video frame, so the reconstruction quality directly affects the prediction efficiency, and the lower reconstruction quality will lead to an increase in the corresponding residual information encoding overhead

When the residual coding overhead is limited, the quality of the reconstructed frame will be significantly reduced, eventually causing serious inter-frame error propagation problems

In addition, the complex network structure and prediction compensation framework make the model design of variable bit rate coding more complicated, so these methods train multiple models independently for different bit rate points

Images