Super-resolution reconstruction method for real-time video session service

Abstract

The invention provides a super-resolution reconstruction method for a real-time video session service, and relates to the technical field of digital image processing. According to the invention, the method comprises the steps: redesigning each super-resolution module, enabling a feature extraction module to adopt coarse-to-fine feature extraction and adopt a residual idea to accelerate the feature extraction speed, introducing deformable convolution into the video super-resolution reconstruction method, and through the idea of a recurrent neural network, dynamically optimizing a frame difference learning module to obtain an optimal alignment parameter; employing the optimal parameter for guiding deformable convolution to carry out alignment operation; designing a feature fusion network for enhancing correlation to carry out feature fusion of adjacent frames, and finally designing a reconstruction module by adopting an information distillation thought, designing an up-sampling reconstruction module, extracting more edge and texture features by using an information distillation block, and adding the extracted edge and texture features with an up-sampled reference frame to generate a final high-resolution video frame. The method is high in reconstruction speed and good in reconstruction quality.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a super-resolution reconstruction method oriented to real-time video session services. Background technique [0002] In recent years, with the rapid development of mobile communications and smart devices, especially under the epidemic this year, applications such as real-time video conferencing and distance education have become an indispensable part of most people's lives. It is precisely because of the interaction between people through video that people's requirements for video quality are becoming higher and higher. So how to acquire, transmit and display high-resolution video with better quality becomes more and more important. However, under the premise of satisfying both video service requirements and video quality, there are two challenges. [0003] On the one hand, from the perspective of hardware, video capture devices have been developed very well at...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is to provide a super-resolution reconstruction method for real-time video conversation services to solve the problem of adaptation and the problem of video quality degradation caused by transmission, and the reconstruction speed is fast. The speed meets real-time performance, and the reconstruction quality is good

Images