Video Coding Partition Prediction With Weighted Motion Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently processing increasing amounts of digital video data, particularly in inter prediction functions for frame prediction, leading to suboptimal encoding and decoding efficiency and speed.
Innovation Solution
Implementing a video encoder and decoder that utilize weighted motion vectors for partition prediction, incorporating an inter predictor and intra predictor to enhance the encoding and decoding processes, including circuitry and memory components to manage motion vector storage and weighted sample processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional inter prediction is used to build prediction of current frame based on reference frame, then the basic encoding function is maintained, but encoding efficiency is insufficient for increasing amounts of digital video data
Solution Approach 1:
The current block is divided into multiple partitions (e.g., 2N×2N, 2N×N, N×2N, N×N, or asymmetric partitions like 16×8, 8×16, 32×16, 16×32). Each partition can be independently predicted using different motion vectors, allowing finer-grained motion compensation that better adapts to local motion variations within the block, thereby improving encoding efficiency for complex video content.
Solution Approach 2:
Different motion vectors are applied to different partitions within the same current block. The inter predictor selects motion vectors from candidate lists (merge mode) or performs motion estimation independently for each partition, enabling local optimization of prediction accuracy according to local motion characteristics, thus improving overall encoding efficiency.
2Measurement precision
If motion vectors from different partitions are used to predict samples, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The system dynamically selects between different prediction modes (merge mode with candidate motion vectors, or independent motion estimation) and partition configurations based on content characteristics. The inter predictor adapts the level of complexity applied to each block, using simpler methods when appropriate and more complex methods only when needed, thus balancing prediction accuracy with computational complexity.
Solution Approach 2:
The system changes parameters such as partition size (2N×2N, 2N×N, N×2N, N×N, asymmetric partitions) and motion vector selection strategy (merge mode candidates vs. full motion estimation) to optimize the trade-off between prediction accuracy and computational complexity for different video content regions.
3Manufacturing precision
If multiple motion vectors are processed and weighted samples are calculated, then encoding precision is improved, but processing time increases
Solution Approach 1:
Motion vector candidates are pre-collected in merge mode from neighboring blocks and reference frames before the actual prediction process. This preliminary preparation of candidate motion vectors reduces the computational burden during encoding, as the inter predictor only needs to evaluate and select from pre-computed candidates rather than performing full motion estimation from scratch, thus reducing processing time while maintaining precision.
Data Source
AI summary
An image encoder or decoder includes circuitry and a memory coupled to the circuitry. The circuitry, in operation, predicts a first set of samples for a first partition of a current picture with one or more motion vectors including a first motion vector and predicts a second set of samples for a first portion of the first partition with one or more motion vectors from a second partition different from the first partition. The samples of the first set of samples of the first portion of the first partition and of the second set of samples of the first portion of the first partition are weighted. A motion vector for the first portion of the first partition is stored which is based on one or both of the first motion vector and the second motion vector. The first partition is encoded or decoded using at least the weighted samples of the first portion of the first partition.