Video Encoding Merge Mode Motion Vector Difference Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding and decoding techniques face inefficiencies in handling increasing video data volumes due to rising resolution, frame rates, and data sizes, necessitating improved compression methods with better encoding efficiency and image quality.
Innovation Solution
The implementation of a video encoding/decoding method and apparatus that applies a merge mode with motion vector difference to combined inter/intra prediction (CIIP) and geometric partitioning modes, determining whether to use CIIP or geometric partitioning, deriving and correcting motion vectors, and generating prediction blocks based on these modes to enhance encoding/decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If video data is compressed using existing techniques (H.264/AVC, HEVC), then storage and transmission are feasible, but encoding efficiency and image quality are insufficient for increasing video size, resolution, and frame rate
Solution Approach 1:
The patent applies geometric partitioning mode that divides a coding unit into multiple regions (e.g., triangular, quadrangular, or polygonal partitions) and performs separate prediction for each region. This segmentation allows more precise modeling of different motion patterns within the same block, improving encoding efficiency for high-resolution and high-frame-rate video content without requiring proportionally increased bitrate
Solution Approach 2:
The patent introduces dynamic motion vector correction using merge mode with motion vector difference (MMVD). Instead of using static motion compensation, the system dynamically adjusts motion vectors by adding differences derived from neighboring blocks, allowing the prediction to adapt to local motion variations. This dynamic approach improves prediction accuracy for complex motion patterns in modern video content
2Measurement precision
If motion vectors are used for inter prediction in CIIP or geometric partitioning mode, then prediction accuracy improves, but additional correction steps increase processing complexity
Solution Approach 1:
The patent performs preliminary motion vector derivation using merge mode, selecting motion vectors from candidate lists based on neighboring blocks. This preliminary action establishes a baseline prediction that is then refined through MMVD correction, avoiding the need to compute complex motion vectors from scratch while still achieving high accuracy
Solution Approach 2:
The patent introduces motion vector difference as an intermediary between the base motion vector from merge mode and the final corrected motion vector. This intermediary representation allows precise motion correction with reduced complexity, as the difference values are typically smaller and require fewer bits than complete motion vectors, balancing accuracy and processing requirements
3Adaptability or versatility
If combined inter/intra prediction mode is applied to generate prediction blocks, then prediction flexibility improves, but motion vector correction becomes necessary adding computational overhead
Solution Approach 1:
The patent applies different prediction strategies to different parts of the prediction block through geometric partitioning. Each partitioned region can use its own motion vector and prediction parameters, allowing local adaptation to different motion characteristics within the same coding unit. This local quality approach provides prediction flexibility while managing computational overhead through selective application of complex modes
Data Source
AI summary
A method and an apparatus for encoding/decoding a video are provided. A method for decoding a video according to the present disclosure includes determining whether to apply a combined inter/intra prediction (CIIP) mode or a geometric partitioning mode. The method for decoding the video also includes deriving a motion vector from a merge candidate list. The method for decoding the video also includes correcting the motion vector by using motion vector difference information. The method for decoding the video also includes generating a prediction block of a current block based on a corrected motion vector.