A fast cu partition decision method for h.266/vvc

Abstract

The present invention proposes a fast CU partition decision method for H.266 / VVC, the steps of which are as follows: firstly, use a traditional coding method to code a video sequence, and record features related to the category of CU partition during the coding process, And use the improved F-score feature selection method to calculate the score value of the feature, and use the feature with a higher score value as a feature subset; secondly, construct an improved directed acyclic graph DAG according to the categories and feature subsets divided by CU -SVM classifier model; finally, input the texture complexity, direction complexity and quantization step size of the CU to be divided into the improved directed acyclic graph DAG-SVM classifier model to predict the best division category of the CU. The present invention combines the improved F-score feature selection method with the improved Directed Acyclic Graph Support Vector Machine (DAG-SVM) model to predict the optimal CU partition in advance, reduces computational complexity, and saves coding time.

Description

Technical field [0001] The present invention relates to the field of image processing, in particular to a rapid Cu partition decision making method for H.266 / VVC. Background technique [0002] With the rise of video applications such as 4K / 8K UHD (UHD), video data has increased, and the encoding technology has been put forward higher requirements. The Motion Image Expert Group (MPEG) and Video Coding Expert Group (VCEG) established a joint video exploration group (JVET), responsible for the development of next-generation video coding standard --H.266 / VVC project. As of 2020, JVET released the latest version of the video test model (VTM8.0) of H.266 / VVC. Compared to HM HM HM, H.266 / VVC maintains the subjective visual quality and increases coding efficiency by approximately 40%. [0003] H.266 / VVC uses a hybrid encoding technology framework, image division from a single, fixed division, continuously moving towards a variety, flexible division structure, and can more eff...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the technical problem of high computational complexity of H.266 / VVC encoding, the present invention proposes a fast CU partition decision method for H.266 / VVC, through the improved F-score feature selection method and the improved directed The combination of ring graph support vector machine DAG-SVM model can predict the best CU partition in advance, reducing the computational complexity and saving coding time

Images