Video Encoding With Adaptive Weighting Factors for Multi-Hypothesis Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques, such as multi-hypothesis prediction (MHP), are limited by a restricted number of possible weighting factor values for inter prediction, leading to reduced coding efficiency, especially when applied to smaller prediction blocks.
Innovation Solution
An inter prediction procedure that extends the number of possible weighting factor values for multi-hypothesis prediction, with restrictions based on prediction block size, and encodes the absolute value and sign of the weighting factor separately to minimize overhead bits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the number of weighting factor values is restricted to a small set, then the complexity of encoding and decoding is reduced, but the coding efficiency deteriorates
Solution Approach 1:
The weighting factor selection process is segmented into multiple stages: first selecting a candidate set based on block size, then selecting final weighting factors from that candidate set. This segmentation allows the system to maintain low complexity for small blocks while enabling access to more weighting factor options for larger blocks, thereby resolving the contradiction between complexity and coding efficiency.
Solution Approach 2:
The candidate set of weighting factors is made dynamic and adaptive based on the prediction block size. Different block sizes trigger different candidate sets, allowing the system to optimize between complexity and efficiency dynamically. Small blocks use a restricted set for simplicity, while large blocks can utilize an expanded set for improved coding efficiency.
2Productivity
If more candidates for weighting factors are allowed, then the coding efficiency is improved, but the overhead bits increase
Solution Approach 1:
The weighting factor candidate sets are segmented according to block size categories. By dividing the weighting factor selection into size-based groups, the patent transmits only the necessary information for each block type, reducing overall overhead bits while maintaining coding efficiency for each specific block size.
Solution Approach 2:
Different block sizes receive different candidate sets tailored to their specific needs. This local optimization ensures that each block type gets the appropriate number of weighting factor candidates, avoiding the transmission of unnecessary overhead bits for blocks that don't require extensive candidate sets.
3Loss of information
If a restricted number of weighting factor values is used, then the overhead bits are minimized, but the compression performance deteriorates
Solution Approach 1:
The system dynamically adjusts the number of weighting factor candidates based on prediction block size. This dynamic adaptation allows the system to minimize overhead bits for small blocks while providing enhanced compression performance for larger blocks that benefit from more weighting factor options, effectively resolving the contradiction between overhead and performance.
Solution Approach 2:
The patent changes the parameter of weighting factor candidate count based on the prediction block size parameter. By making this parameter change adaptive, the system optimizes the balance between overhead bits and compression performance for different block types, improving overall compression performance without excessive overhead.
Data Source
AI summary
method of encoding by an encoder is provided. The method includes receiving, by at least one processor, a set of frames including a reference frame and a current frame. The method includes performing, by the at least one processor, a multi-hypothesis prediction (MHP) procedure for a coding unit (CU) located in the current frame based on a search block in the reference frame. In response to a size of the search block in the reference frame meeting a threshold size, the method includes selecting, by the at least one processor, a first weighting factor from a first set of more than two weighting factors associated with the MHP procedure.