Memory bandwidth constrained template-based operations

By introducing template-based decoding constraints and adaptive merge candidate reordering at the video block level, the determination process of motion vector prediction candidates is optimized, and the problem of inefficient video decoding under the constraints of memory bandwidth is solved, and more efficient Video block decoding.

CN119234418APending Publication Date: 2024-12-31INTERDIGITAL CE PATENT HOLDINGS SAS
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380039410.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-04-08
Filing Date
2023-04-07
Publication Date
2024-12-31

AI Technical Summary

Technical Problem

Existing video decoding systems are difficult to efficiently perform template-based operations under the constraints of memory bandwidth, especially in the determination and merging of motion vector prediction (MVP) candidates, resulting in inefficiency.

Method used

The validity of MVP candidates is determined by introducing template-based coding constraints, such as region constraints and reference picture constraints, and performing adaptive merge candidate reordering (ARMC) to optimize memory bandwidth usage and correct motion vectors by introducing template-based coding constraints, such as region constraints and reference picture constraints, and to perform adaptive merge candidate reordering (ARMC). Candidate subsets to improve decoding efficiency.

Benefits of technology

It effectively reduces the memory bandwidth requirement, improves the decoding efficiency of video blocks, reduces template costs, and improves the performance of video decoding and encoding processes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119234418A_ABST
    Figure CN119234418A_ABST
Patent Text Reader

Abstract

Systems, methods, and instrumentalities are disclosed for performing memory bandwidth constrained template-based operations. For example, a device (e.g., a video decoding device, a video encoding device) may obtain a motion vector prediction (MVP) candidate associated with a video block. The MVP candidate may include a first MVP candidate. The device may determine, for example, whether the first MVP candidate can be used for template-based coding based on template-based coding constraints (e.g., region constraints, reference picture constraints, and / or coding tree unit size constraints, etc. The device may perform a template-based function associated with the first MVP candidate associated with the video block (e.g., based on determining that the first MVP candidate can be used for template-based coding).
Need to check novelty before this filing date? Find Prior Art