ETMVP Motion Prediction for Subblock Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding methods, such as those in the Versatile Video Coding (VVC) standard, face inefficiencies in temporal motion vector prediction, particularly in handling motion information for subblocks within coding units, leading to suboptimal compression performance.
Innovation Solution
The introduction of Enhanced Temporal Motion Vector Prediction (ETMVP) that includes determining an ETMVP mode flag and index value for a current coding unit (CU), constructing an enhanced temporal candidate list, and using motion information from a target matching CU to improve prediction accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ATMVP technique is used to provide different motion information for each subblock, then motion prediction accuracy is improved, but encoding complexity and bit usage increase
Solution Approach 1:
The patent changes the parameter of motion vector prediction by introducing multiple ETMVP modes (first mode using temporal subblock motion information, second mode using spatial neighboring block motion information) instead of a single ATMVP approach. This allows the system to adaptively select the most efficient prediction method for each coding unit, improving accuracy while managing complexity through conditional mode selection based on CU size and other factors.
2Measurement precision
If motion information is provided for each subblock individually, then prediction quality is improved, but bit rate consumption increases
Solution Approach 1:
The patent applies local quality by differentiating the treatment of motion information based on the characteristics of different coding units. Larger CUs use the first ETMVP mode with temporal subblock information, while smaller CUs use the second ETMVP mode with spatial neighboring block information. This localized adaptation of prediction strategies optimizes prediction quality for each region's specific needs while minimizing overall bit rate consumption.
3Adaptability or versatility
If enhanced temporal candidate list is constructed with multiple modes, then encoding flexibility is improved, but processing time increases
Solution Approach 1:
The patent introduces dynamic mode selection where the ETMVP mode is determined adaptively based on coding unit size, prediction mode, and other contextual factors. The system dynamically switches between first ETMVP mode (temporal-based) and second ETMVP mode (spatial-based) rather than using a static approach. This dynamic adaptation provides encoding flexibility while controlling processing time by avoiding unnecessary complexity for all coding units.
Data Source
AI summary
Methods, apparatuses, devices and storage media for decoding and encoding are provided. In one aspect, a decoding method includes: obtaining a value of an Enhanced Temporal Motion Vector Prediction (ETMVP) mode flag and an ETMVP mode index value for a current coding unit (CU); when the value of the ETMVP mode flag is a first value, determining coordinate information of an initial matching CU of the current CU according to coordinate information of the current CU; constructing an enhanced temporal candidate list based on the coordinate information of the initial matching CU; determining coordinate information of a target matching CU according to the ETMVP mode index value; and determining motion information of a subblock at a corresponding position in the current CU according to the motion information of the target matching CU.