Encoder Motion Vector Segmentation for Accurate Inter Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like H.265/HEVC face challenges in achieving efficient inter prediction due to limitations in deriving accurate motion vectors, which affect encoding efficiency.
Innovation Solution
An encoder and decoder system that determines a base motion vector and encodes/decodes a delta motion vector representing a direction including a diagonal direction and distance from the base motion vector, improving inter prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion vector derivation methods are used in H.265/HEVC, then the encoding process is simple, but the motion vector accuracy is insufficient
Solution Approach 1:
The motion vector is segmented into a base motion vector (derived from neighboring blocks) and a delta motion vector (representing direction and distance adjustments). This segmentation allows the encoder to separately handle the coarse motion estimation and fine adjustments, improving accuracy without requiring complete re-estimation of the entire motion vector.
Solution Approach 2:
The patent introduces a new dimensional representation for the delta motion vector by encoding it as a combination of direction (from a set including diagonal directions) and distance from the base motion vector. This dimensional transformation allows more precise motion representation using compact syntax elements.
2Productivity
If higher-accuracy motion vectors are derived using delta motion vectors with diagonal directions, then inter prediction efficiency improves, but the encoding complexity increases
Solution Approach 1:
The patent applies local quality by allowing diagonal directions and variable distance adjustments in the delta motion vector based on the specific characteristics of each current block. This enables adaptive refinement of motion vectors where needed while maintaining simplicity in regions where conventional methods suffice.
Solution Approach 2:
The patent changes the parameters used to represent motion vector differences by introducing direction indices (selecting from multiple directions including diagonals) and distance values. This parameter transformation enables more efficient representation of motion variations with compact syntax elements that improve prediction accuracy.
Data Source
AI summary
An encoder, includes: circuitry; and memory. Using the memory, the circuitry: in inter prediction for a current block, determines a base motion vector, and writes, in an encoded signal, a delta motion vector representing (i) one direction among a plurality of directions including a diagonal direction and (ii) a distance from the base motion vector; and encodes the current block using the delta motion vector and the base motion vector as a motion vector of the current block.


