Super-resolution reconstruction method and device for adaptive streaming media and server

Abstract

The invention discloses a super-resolution reconstruction method and device for adaptive streaming media and a server, and the method comprises the steps: in the playing process of the streaming media, when the first resolution of a video stream downloaded by terminal equipment does not reach a target super-resolution, the terminal equipment carries out the super-resolution reconstruction of the video stream according to the first resolution and the target super-resolution; obtaining a corresponding super-resolution neural network from a server; wherein the super-resolution neural network is obtained by performing model training in advance by taking the video stream of the first resolution of the streaming media as input and taking the video stream of the target super-resolution of the streaming media as a learning target; the target super-resolution is greater than the first resolution; and the terminal device reconstructs the downloaded video stream into a video stream with the target super-resolution by using the super-resolution neural network, and plays the streaming media based on the video stream obtained by reconstruction. According to the invention, the video image quality of the adaptive streaming media can be improved.

Description

technical field [0001] The present invention relates to computer application technology, in particular to a super-resolution reconstruction method, device and server for adaptive streaming media. Background technique [0002] At present, in the field of video playback, dynamic adaptive streaming media transmission technology (such as smooth stream, HLS and DASH standards, etc.) has been widely used. This technology can select video streams with different bit rates and resolutions according to different network conditions. To ensure the smoothness of video playback. [0003] During the process of implementing the present invention, the inventor found that the existing dynamic adaptive code stream playing scheme cannot ensure the video quality while ensuring the smoothness of video playback. The main reason for this problem is that in the existing dynamic adaptive code stream playback scheme, in order to ensure the smoothness of video playback, the real-time network status wi...

AI Technical Summary

Problems solved by technology

[0003] In the process of implementing the present invention, the inventor found that the existing dynamic adaptive code stream playback scheme cannot guarantee the video quality while ensuring the smoothness of video playback.

The main reason for this problem is: the existing dynamic adaptive code stream playback scheme, in order to ensure the fluency of video playback, will consider the real-time network status and control the resolution of the video stream download, for example, when the network bandwidth is poor , often choose low-resolution video playback. In this way, the resolution of the video stream actually downloaded by the user often cannot reach the user's target playback resolution, so that the quality of the played video cannot meet the user's needs, and even after switching to a lower resolution , you will obviously feel that the picture quality is reduced, resulting in a poor viewing experience for users

Images