Method for improving video error concealment effect by utilizing generative network

Abstract

The invention discloses a method for improving the video error concealment effect by utilizing a generative network. According to the method disclosed by the invention, on the basis of an error concealment technology based on the H.264 standard, a CNN structural network G is constructed; a 16*16 repaired macro block is subjected to G processing; therefore, the purpose of improving the PSNR is realized; in addition, an input image is directly mapped before an output image in the G network, and added with the network fitting residual, so that the final output is obtained; notably, the method isonly for the H.264 standard, and not suitable for other video coding standards; by means of the method disclosed by the invention, the network convergence is relatively rapid; a training result is obtained rapidly and better; a deep neural network is a high-efficiency tool in image processing; on the basis of an inter-frame concealment technology, a neural network is added; therefore, the psnr value of the repaired macro block can be increased; the quality of a whole picture is relatively high; and thus, the video error concealment effect is also better.

Description

technical field [0001] The invention relates to the field of deep neural network and the field of video error concealment based on the H.264 standard, in particular to a CNN structure network and a residual structure. technical background [0002] People's requirements for video quality are constantly improving. With the increase of information volume and the urgent need for high-density storage, video coding and compression technology becomes very important in the transmission process. Video compression technology mainly uses certain algorithms to remove video Redundancy such as time and space can compress the size of the video to a certain extent, but at the cost of a certain amount of information loss. The current mainstream video compression standards are divided into H.26X and MPEG.X series. The former mainly focuses on improving compression performance, while the latter pursues the addition of various functions at the application level. At present, the H.264 compress...

AI Technical Summary

Problems solved by technology

In addition, the image pixels obtained by the traditional technology using MSE (mean square error) as the loss function are too smooth, and the local texture details are less. Although the low-quality images on PSNR have been greatly improved, the visual experience There is no expected difference, but an adversarial identification model is constructed in SRGAN, and the traditional loss function is abandoned and the content loss function and the adversarial loss function are combined into a perceptual loss function, which makes the generated image texture details more, Let people have a stronger feeling in visual perception

Images