An intelligent and fast inter-frame coding method

Abstract

The invention discloses an intelligent fast inter-frame coding method, which includes: considering the division of coding units as a binary classification problem, designing a multi-information fusion convolutional neural network model as a classifier, and realizing the division of coding units in advance Termination; take the texture information and motion information of the coding unit as the input of the classifier, and obtain the input features through the asymmetric convolution kernel feature extraction module respectively; then use the attention-based feature fusion module to obtain the fused features, and finally use the classification module The classification result is obtained according to the fusion feature; by considering the motion information of the coding unit and the confidence degree of the output of the multi-information fusion network as conditions, the early decision-making of the Merge mode is realized. The present invention shortens the inter-frame coding time as much as possible under the condition that the compressed video quality remains unchanged, and effectively reduces the inter-frame coding complexity by making full use of the spatial domain, time domain and sports field information by virtue of the feature extraction and analysis capabilities of deep learning.

Description

Technical field [0001] The present invention relates to deep learning, video encoding, in particular to an intelligent fast inter-frame encoding method. Background [0002] Video is an important carrier of information dissemination. In recent years, with the rapid development of the Internet and multimedia technology, the performance of hardware equipment has been continuously improved, and the application scenarios of high-definition, ultra-high-definition and panoramic video have become more and more extensive. However, this larger volume of video poses great challenges for storage and transmission. [0003] Video compression encoding technology helps to solve the contradiction between the huge amount of video data and the limited transmission bandwidth. Through efficient video compression and coding technology, redundant information in the video can be removed and the need for high transmission bandwidth of the video can be reduced. However, the high computational complexity ...

AI Technical Summary

Problems solved by technology

[0005] Existing methods mainly focus on using the texture information of coding units to reduce the complexity of intra-frame coding, and do not consider the temporal correlation between frames, so these methods are not suitable for reducing the complexity of inter-frame coding

There is currently no work on low-complexity inter-frame coding methods based on deep learning

Images